2019-06-14 10:16:30 +00:00
package wallet
import (
"database/sql"
"database/sql/driver"
"encoding/json"
"errors"
"math/big"
2019-07-15 11:16:07 +00:00
"reflect"
2019-06-14 10:16:30 +00:00
"github.com/ethereum/go-ethereum/common"
"github.com/ethereum/go-ethereum/core/types"
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
"github.com/ethereum/go-ethereum/log"
2019-06-14 10:16:30 +00:00
)
// DBHeader fields from header that are stored in database.
type DBHeader struct {
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
Number * big . Int
Hash common . Hash
Timestamp uint64
Erc20Transfer * Transfer
Network uint64
Address common . Address
2019-06-14 10:16:30 +00:00
// Head is true if the block was a head at the time it was pulled from chain.
Head bool
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
// Loaded is true if trasfers from this block has been already fetched
Loaded bool
2019-06-14 10:16:30 +00:00
}
func toDBHeader ( header * types . Header ) * DBHeader {
return & DBHeader {
2019-07-10 09:08:43 +00:00
Hash : header . Hash ( ) ,
Number : header . Number ,
Timestamp : header . Time ,
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
Loaded : false ,
2019-06-14 10:16:30 +00:00
}
}
func toHead ( header * types . Header ) * DBHeader {
2019-07-10 09:08:43 +00:00
dbheader := toDBHeader ( header )
dbheader . Head = true
return dbheader
2019-06-14 10:16:30 +00:00
}
// SyncOption is used to specify that application processed transfers for that block.
type SyncOption uint
// SQLBigInt type for storing uint256 in the databse.
// FIXME(dshulyak) SQL big int is max 64 bits. Maybe store as bytes in big endian and hope
// that lexographical sorting will work.
type SQLBigInt big . Int
// Scan implements interface.
func ( i * SQLBigInt ) Scan ( value interface { } ) error {
val , ok := value . ( int64 )
if ! ok {
return errors . New ( "not an integer" )
}
( * big . Int ) ( i ) . SetInt64 ( val )
return nil
}
// Value implements interface.
func ( i * SQLBigInt ) Value ( ) ( driver . Value , error ) {
if ! ( * big . Int ) ( i ) . IsInt64 ( ) {
2019-07-10 09:08:43 +00:00
return nil , errors . New ( "not an int64" )
2019-06-14 10:16:30 +00:00
}
return ( * big . Int ) ( i ) . Int64 ( ) , nil
}
// JSONBlob type for marshaling/unmarshaling inner type to json.
type JSONBlob struct {
data interface { }
}
// Scan implements interface.
func ( blob * JSONBlob ) Scan ( value interface { } ) error {
2019-07-15 11:16:07 +00:00
if value == nil || reflect . ValueOf ( blob . data ) . IsNil ( ) {
return nil
}
2019-06-14 10:16:30 +00:00
bytes , ok := value . ( [ ] byte )
if ! ok {
return errors . New ( "not a byte slice" )
}
if len ( bytes ) == 0 {
return nil
}
err := json . Unmarshal ( bytes , blob . data )
return err
}
// Value implements interface.
func ( blob * JSONBlob ) Value ( ) ( driver . Value , error ) {
2019-07-15 11:16:07 +00:00
if blob . data == nil || reflect . ValueOf ( blob . data ) . IsNil ( ) {
return nil , nil
}
2019-06-14 10:16:30 +00:00
return json . Marshal ( blob . data )
}
2019-08-08 08:15:07 +00:00
func NewDB ( db * sql . DB , network uint64 ) * Database {
return & Database { db : db , network : network }
2019-07-25 05:35:09 +00:00
}
2019-06-14 10:16:30 +00:00
// Database sql wrapper for operations with wallet objects.
type Database struct {
2019-08-08 08:15:07 +00:00
db * sql . DB
network uint64
2019-06-14 10:16:30 +00:00
}
// Close closes database.
func ( db Database ) Close ( ) error {
return db . db . Close ( )
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
func ( db Database ) ProcessBlocks ( account common . Address , from * big . Int , to * big . Int , headers [ ] * DBHeader ) ( err error ) {
var (
tx * sql . Tx
)
tx , err = db . db . Begin ( )
if err != nil {
return err
}
defer func ( ) {
if err == nil {
err = tx . Commit ( )
return
}
_ = tx . Rollback ( )
} ( )
err = insertBlocksWithTransactions ( tx , account , db . network , headers )
if err != nil {
return
}
err = insertRange ( tx , account , db . network , from , to )
if err != nil {
return
}
return
}
func ( db Database ) SaveBlocks ( account common . Address , headers [ ] * DBHeader ) ( err error ) {
var (
tx * sql . Tx
)
tx , err = db . db . Begin ( )
if err != nil {
return err
}
defer func ( ) {
if err == nil {
err = tx . Commit ( )
return
}
_ = tx . Rollback ( )
} ( )
err = insertBlocksWithTransactions ( tx , account , db . network , headers )
if err != nil {
return
}
return
}
2019-06-14 10:16:30 +00:00
// ProcessTranfers atomically adds/removes blocks and adds new tranfers.
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
func ( db Database ) ProcessTranfers ( transfers [ ] Transfer , removed [ ] * DBHeader ) ( err error ) {
2019-06-14 10:16:30 +00:00
var (
tx * sql . Tx
)
tx , err = db . db . Begin ( )
if err != nil {
return err
}
defer func ( ) {
if err == nil {
err = tx . Commit ( )
return
}
_ = tx . Rollback ( )
} ( )
err = deleteHeaders ( tx , removed )
if err != nil {
return
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
err = updateOrInsertTransfers ( tx , db . network , transfers )
2019-06-14 10:16:30 +00:00
if err != nil {
return
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
return
}
// SaveTranfers
func ( db Database ) SaveTranfers ( address common . Address , transfers [ ] Transfer , blocks [ ] * big . Int ) ( err error ) {
var (
tx * sql . Tx
)
tx , err = db . db . Begin ( )
if err != nil {
return err
}
defer func ( ) {
if err == nil {
err = tx . Commit ( )
return
}
_ = tx . Rollback ( )
} ( )
err = updateOrInsertTransfers ( tx , db . network , transfers )
2019-06-14 10:16:30 +00:00
if err != nil {
return
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
err = markBlocksAsLoaded ( tx , address , db . network , blocks )
if err != nil {
return
}
2019-06-14 10:16:30 +00:00
return
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
// GetTransfersInRange loads transfers for a given address between two blocks.
func ( db * Database ) GetTransfersInRange ( address common . Address , start , end * big . Int ) ( rst [ ] Transfer , err error ) {
query := newTransfersQuery ( ) . FilterNetwork ( db . network ) . FilterAddress ( address ) . FilterStart ( start ) . FilterEnd ( end ) . FilterLoaded ( 1 )
2019-06-14 10:16:30 +00:00
rows , err := db . db . Query ( query . String ( ) , query . Args ( ) ... )
if err != nil {
return
}
defer rows . Close ( )
return query . Scan ( rows )
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
// GetTransfersByAddress loads transfers for a given address between two blocks.
2020-02-07 10:26:00 +00:00
func ( db * Database ) GetTransfersByAddress ( address common . Address , toBlock * big . Int , limit int64 ) ( rst [ ] Transfer , err error ) {
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
query := newTransfersQuery ( ) .
FilterNetwork ( db . network ) .
FilterAddress ( address ) .
2020-02-07 10:26:00 +00:00
FilterEnd ( toBlock ) .
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
FilterLoaded ( 1 ) .
Limit ( limit )
2019-06-14 10:16:30 +00:00
rows , err := db . db . Query ( query . String ( ) , query . Args ( ) ... )
if err != nil {
return
}
defer rows . Close ( )
return query . Scan ( rows )
}
2020-02-07 10:26:00 +00:00
// GetBlocksByAddress loads blocks for a given address.
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
func ( db * Database ) GetBlocksByAddress ( address common . Address , limit int ) ( rst [ ] * big . Int , err error ) {
query := ` SELECT blk_number FROM blocks
WHERE address = ? AND network_id = ? AND loaded = 0
ORDER BY blk_number DESC
LIMIT ? `
rows , err := db . db . Query ( query , address , db . network , limit )
2019-06-14 10:16:30 +00:00
if err != nil {
return
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
defer rows . Close ( )
for rows . Next ( ) {
block := & big . Int { }
err = rows . Scan ( ( * SQLBigInt ) ( block ) )
if err != nil {
return nil , err
}
rst = append ( rst , block )
}
return rst , nil
}
func ( db * Database ) RemoveBlockWithTransfer ( address common . Address , block * big . Int ) error {
query := ` DELETE FROM blocks
WHERE address = ?
AND blk_number = ?
AND network_id = ? `
_ , err := db . db . Exec ( query , address , ( * SQLBigInt ) ( block ) , db . network )
if err != nil {
return err
}
return nil
}
func ( db * Database ) GetLastBlockByAddress ( address common . Address , limit int ) ( rst * big . Int , err error ) {
query := ` SELECT * FROM
( SELECT blk_number FROM blocks WHERE address = ? AND network_id = ? ORDER BY blk_number DESC LIMIT ? )
ORDER BY blk_number LIMIT 1 `
rows , err := db . db . Query ( query , address , db . network , limit )
2019-06-14 10:16:30 +00:00
if err != nil {
return
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
defer rows . Close ( )
if rows . Next ( ) {
block := & big . Int { }
err = rows . Scan ( ( * SQLBigInt ) ( block ) )
if err != nil {
return nil , err
2019-06-14 10:16:30 +00:00
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
return block , nil
}
return nil , nil
}
func ( db * Database ) GetLastSavedBlock ( ) ( rst * DBHeader , err error ) {
query := ` SELECT blk_number , blk_hash
FROM blocks
WHERE network_id = ?
ORDER BY blk_number DESC LIMIT 1 `
rows , err := db . db . Query ( query , db . network )
if err != nil {
return
}
defer rows . Close ( )
if rows . Next ( ) {
header := & DBHeader { Hash : common . Hash { } , Number : new ( big . Int ) }
err = rows . Scan ( ( * SQLBigInt ) ( header . Number ) , & header . Hash )
2019-06-14 10:16:30 +00:00
if err != nil {
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
return nil , err
2019-06-14 10:16:30 +00:00
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
return header , nil
2019-06-14 10:16:30 +00:00
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
return nil , nil
2019-06-14 10:16:30 +00:00
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
func ( db * Database ) GetBlocks ( ) ( rst [ ] * DBHeader , err error ) {
query := ` SELECT blk_number, blk_hash, address FROM blocks `
rows , err := db . db . Query ( query , db . network )
2019-06-14 10:16:30 +00:00
if err != nil {
return
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
defer rows . Close ( )
rst = [ ] * DBHeader { }
for rows . Next ( ) {
header := & DBHeader { Hash : common . Hash { } , Number : new ( big . Int ) }
err = rows . Scan ( ( * SQLBigInt ) ( header . Number ) , & header . Hash , & header . Address )
if err != nil {
return nil , err
}
rst = append ( rst , header )
}
return rst , nil
}
func ( db * Database ) GetLastSavedBlockBefore ( block * big . Int ) ( rst * DBHeader , err error ) {
query := ` SELECT blk_number , blk_hash
FROM blocks
WHERE network_id = ? AND blk_number < ?
ORDER BY blk_number DESC LIMIT 1 `
rows , err := db . db . Query ( query , db . network , ( * SQLBigInt ) ( block ) )
2019-06-14 10:16:30 +00:00
if err != nil {
return
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
defer rows . Close ( )
if rows . Next ( ) {
header := & DBHeader { Hash : common . Hash { } , Number : new ( big . Int ) }
err = rows . Scan ( ( * SQLBigInt ) ( header . Number ) , & header . Hash )
if err != nil {
return nil , err
}
return header , nil
}
return nil , nil
}
func ( db * Database ) GetFirstKnownBlock ( address common . Address ) ( rst * big . Int , err error ) {
query := ` SELECT blk_from FROM blocks_ranges
WHERE address = ?
AND network_id = ?
ORDER BY blk_from
LIMIT 1 `
rows , err := db . db . Query ( query , address , db . network )
if err != nil {
return
}
defer rows . Close ( )
if rows . Next ( ) {
block := & big . Int { }
err = rows . Scan ( ( * SQLBigInt ) ( block ) )
if err != nil {
return nil , err
}
return block , nil
}
return nil , nil
}
func ( db * Database ) GetLastKnownBlockByAddress ( address common . Address ) ( rst * big . Int , err error ) {
query := ` SELECT blk_to FROM blocks_ranges
WHERE address = ?
AND network_id = ?
ORDER BY blk_to DESC
LIMIT 1 `
rows , err := db . db . Query ( query , address , db . network )
if err != nil {
return
}
defer rows . Close ( )
if rows . Next ( ) {
block := & big . Int { }
err = rows . Scan ( ( * SQLBigInt ) ( block ) )
if err != nil {
return nil , err
}
return block , nil
}
return nil , nil
}
func ( db * Database ) GetLastKnownBlockByAddresses ( addresses [ ] common . Address ) ( map [ common . Address ] * big . Int , [ ] common . Address , error ) {
res := map [ common . Address ] * big . Int { }
accountsWithoutHistory := [ ] common . Address { }
for _ , address := range addresses {
block , err := db . GetLastKnownBlockByAddress ( address )
if err != nil {
log . Info ( "Can't get last block" , "error" , err )
return nil , nil , err
}
if block != nil {
res [ address ] = block
2019-06-14 10:16:30 +00:00
} else {
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
accountsWithoutHistory = append ( accountsWithoutHistory , address )
2019-06-14 10:16:30 +00:00
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
}
return res , accountsWithoutHistory , nil
}
// GetTransfers load transfers transfer betweeen two blocks.
func ( db * Database ) GetTransfers ( start , end * big . Int ) ( rst [ ] Transfer , err error ) {
query := newTransfersQuery ( ) . FilterNetwork ( db . network ) . FilterStart ( start ) . FilterEnd ( end ) . FilterLoaded ( 1 )
rows , err := db . db . Query ( query . String ( ) , query . Args ( ) ... )
2019-06-14 10:16:30 +00:00
if err != nil {
return
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
defer rows . Close ( )
return query . Scan ( rows )
2019-06-14 10:16:30 +00:00
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
func ( db * Database ) GetPreloadedTransactions ( address common . Address , blockHash common . Hash ) ( rst [ ] Transfer , err error ) {
query := newTransfersQuery ( ) .
FilterNetwork ( db . network ) .
FilterAddress ( address ) .
FilterBlockHash ( blockHash ) .
FilterLoaded ( 0 )
rows , err := db . db . Query ( query . String ( ) , query . Args ( ) ... )
2019-06-14 10:16:30 +00:00
if err != nil {
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
return
2019-06-14 10:16:30 +00:00
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
defer rows . Close ( )
return query . Scan ( rows )
2019-06-14 10:16:30 +00:00
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
func ( db * Database ) GetTransactionsLog ( address common . Address , transactionHash common . Hash ) ( * types . Log , error ) {
l := & types . Log { }
err := db . db . QueryRow ( "SELECT log FROM transfers WHERE network_id = ? AND address = ? AND hash = ?" ,
db . network , address , transactionHash ) .
Scan ( & JSONBlob { l } )
2019-06-14 10:16:30 +00:00
if err == nil {
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
return l , nil
2019-06-14 10:16:30 +00:00
}
if err == sql . ErrNoRows {
return nil , nil
}
return nil , err
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
// SaveHeaders stores a list of headers atomically.
func ( db * Database ) SaveHeaders ( headers [ ] * types . Header , address common . Address ) ( err error ) {
var (
tx * sql . Tx
insert * sql . Stmt
)
tx , err = db . db . Begin ( )
if err != nil {
return
2019-06-14 10:16:30 +00:00
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
insert , err = tx . Prepare ( "INSERT INTO blocks(network_id, blk_number, blk_hash, address) VALUES (?, ?, ?, ?)" )
if err != nil {
return
2019-06-14 10:16:30 +00:00
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
defer func ( ) {
if err == nil {
err = tx . Commit ( )
} else {
_ = tx . Rollback ( )
}
} ( )
for _ , h := range headers {
_ , err = insert . Exec ( db . network , ( * SQLBigInt ) ( h . Number ) , h . Hash ( ) , address )
if err != nil {
return
}
}
return
2019-06-14 10:16:30 +00:00
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
// GetHeaderByNumber selects header using block number.
func ( db * Database ) GetHeaderByNumber ( number * big . Int ) ( header * DBHeader , err error ) {
2019-06-14 10:16:30 +00:00
header = & DBHeader { Hash : common . Hash { } , Number : new ( big . Int ) }
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
err = db . db . QueryRow ( "SELECT blk_hash, blk_number FROM blocks WHERE blk_number = ? AND network_id = ?" , ( * SQLBigInt ) ( number ) , db . network ) . Scan ( & header . Hash , ( * SQLBigInt ) ( header . Number ) )
2019-06-14 10:16:30 +00:00
if err == nil {
return header , nil
}
if err == sql . ErrNoRows {
return nil , nil
}
return nil , err
}
2019-12-10 17:31:08 +00:00
type Token struct {
Address common . Address ` json:"address" `
Name string ` json:"name" `
Symbol string ` json:"symbol" `
Color string ` json:"color" `
// Decimals defines how divisible the token is. For example, 0 would be
// indivisible, whereas 18 would allow very small amounts of the token
// to be traded.
Decimals uint ` json:"decimals" `
}
func ( db * Database ) GetCustomTokens ( ) ( [ ] * Token , error ) {
rows , err := db . db . Query ( ` SELECT address, name, symbol, decimals, color FROM tokens WHERE network_id = ? ` , db . network )
if err != nil {
return nil , err
}
defer rows . Close ( )
var rst [ ] * Token
for rows . Next ( ) {
token := & Token { }
err := rows . Scan ( & token . Address , & token . Name , & token . Symbol , & token . Decimals , & token . Color )
if err != nil {
return nil , err
}
rst = append ( rst , token )
}
return rst , nil
}
func ( db * Database ) AddCustomToken ( token Token ) error {
insert , err := db . db . Prepare ( "INSERT OR REPLACE INTO TOKENS (network_id, address, name, symbol, decimals, color) VALUES (?, ?, ?, ?, ?, ?)" )
if err != nil {
return err
}
_ , err = insert . Exec ( db . network , token . Address , token . Name , token . Symbol , token . Decimals , token . Color )
return err
}
func ( db * Database ) DeleteCustomToken ( address common . Address ) error {
_ , err := db . db . Exec ( ` DELETE FROM TOKENS WHERE address = ? ` , address )
return err
}
2019-06-14 10:16:30 +00:00
// statementCreator allows to pass transaction or database to use in consumer.
type statementCreator interface {
Prepare ( query string ) ( * sql . Stmt , error )
}
func deleteHeaders ( creator statementCreator , headers [ ] * DBHeader ) error {
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
delete , err := creator . Prepare ( "DELETE FROM blocks WHERE blk_hash = ?" )
if err != nil {
return err
}
deleteTransfers , err := creator . Prepare ( "DELETE FROM transfers WHERE blk_hash = ?" )
2019-06-14 10:16:30 +00:00
if err != nil {
return err
}
for _ , h := range headers {
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
k := h . Hash
log . Debug ( "foo" , "k" , k )
2019-06-14 10:16:30 +00:00
_ , err = delete . Exec ( h . Hash )
if err != nil {
return err
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
_ , err = deleteTransfers . Exec ( h . Hash )
2019-06-14 10:16:30 +00:00
if err != nil {
return err
}
}
return nil
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
func insertBlocksWithTransactions ( creator statementCreator , account common . Address , network uint64 , headers [ ] * DBHeader ) error {
insert , err := creator . Prepare ( "INSERT OR IGNORE INTO blocks(network_id, address, blk_number, blk_hash, loaded) VALUES (?, ?, ?, ?, ?)" )
2019-06-14 10:16:30 +00:00
if err != nil {
return err
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
updateTx , err := creator . Prepare ( ` UPDATE transfers
SET log = ?
WHERE network_id = ? AND address = ? AND hash = ? ` )
2019-06-14 10:16:30 +00:00
if err != nil {
return err
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
insertTx , err := creator . Prepare ( ` INSERT OR IGNORE
INTO transfers ( network_id , address , sender , hash , blk_number , blk_hash , type , timestamp , log , loaded )
VALUES ( ? , ? , ? , ? , ? , ? , ? , ? , ? , 0 ) ` )
2019-06-14 10:16:30 +00:00
if err != nil {
return err
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
for _ , header := range headers {
_ , err = insert . Exec ( network , account , ( * SQLBigInt ) ( header . Number ) , header . Hash , header . Loaded )
if err != nil {
return err
}
if header . Erc20Transfer != nil {
res , err := updateTx . Exec ( & JSONBlob { header . Erc20Transfer . Log } , network , account , header . Erc20Transfer . ID )
2019-06-14 10:16:30 +00:00
if err != nil {
return err
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
affected , err := res . RowsAffected ( )
2019-06-14 10:16:30 +00:00
if err != nil {
return err
}
if affected > 0 {
continue
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
_ , err = insertTx . Exec ( network , account , account , header . Erc20Transfer . ID , ( * SQLBigInt ) ( header . Number ) , header . Hash , erc20Transfer , header . Erc20Transfer . Timestamp , & JSONBlob { header . Erc20Transfer . Log } )
2019-06-14 10:16:30 +00:00
if err != nil {
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
log . Error ( "error saving erc20transfer" , "err" , err )
2019-06-14 10:16:30 +00:00
return err
}
}
}
return nil
}
status-im/status-react#9203 Faster tx fetching with less request
*** How it worked before this PR on multiaccount creation:
- On multiacc creation we scanned chain for eth and erc20 transfers. For
each address of a new empty multiaccount this scan required
1. two `eth_getBalance` requests to find out that there is no any
balance change between zero and the last block, for eth transfers
2. and `chain-size/100000` (currently ~100) `eth_getLogs` requests,
for erc20 transfers
- For some reason we scanned an address of the chat account as well, and
also accounts were not deduplicated. So even for an empty multiacc we
scanned chain twice for each chat and main wallet addresses, in result
app had to execute about 400 requests.
- As mentioned above, `eth_getBalance` requests were used to check if
there were any eth transfers, and that caused empty history in case
if user already used all available eth (so that both zero and latest
blocks show 0 eth for an address). There might have been transactions
but we wouldn't fetch/show them.
- There was no upper limit for the number of rpc requests during the
scan, so it could require indefinite number of requests; the scanning
algorithm was written so that we persisted the whole history of
transactions or tried to scan form the beginning again in case of
failure, giving up only after 10 minutes of failures. In result
addresses with sufficient number of transactions would never be fully
scanned and during these 10 minutes app could use gigabytes of
internet data.
- Failures were caused by `eth_getBlockByNumber`/`eth_getBlockByHash`
requests. These requests return significantly bigger responses than
`eth_getBalance`/`eth_transactionsCount` and it is likely that
execution of thousands of them in parallel caused failures for
accounts with hundreds of transactions. Even for an account with 12k
we could successfully determine blocks with transaction in a few
minutes using `eth_getBalance` requests, but `eth_getBlock...`
couldn't be processed for this acc.
- There was no caching for for `eth_getBalance` requests, and this
caused in average 3-4 times more such requests than is needed.
*** How it works now on multiaccount creation:
- On multiacc creation we scan chain for last ~30 eth transactions and
then check erc20 in the range where these eth transactions were found.
For an empty address in multiacc this means:
1. two `eth_getBalance` transactions to determine that there was no
balance change between zero and the last block; two
`eth_transactionsCount` requests to determine there are no outgoing
transactions for this address; total 4 requests for eth transfers
2. 20 `eth_getLogs` for erc20 transfers. This number can be lowered,
but that's not a big deal
- Deduplication of addresses is added and also we don't scan chat
account, so a new multiacc requires ~25 (we also request latest block
number and probably execute a few other calls) request to determine
that multiacc is empty (comparing to ~400 before)
- In case if address contains transactions we:
1. determine the range which contains 20-25 outgoing eth/erc20
transactions. This usually requires up to 10 `eth_transactionCount`
requests
2. then we scan chain for eth transfers using `eth_getBalance` and
`eth_transactionCount` (for double checking zero balances)
3. we make sure that we do not scan db for more than 30 blocks with
transfers. That's important for accounts with mostly incoming
transactions, because the range found on the first step might
contain any number of incoming transfers, but only 20-25 outgoing
transactions
4. when we found ~30 blocks in a given range, we update initial
range `from` block using the oldest found block
5. and now we scan db for erc20transfers using `eth_getLogs`
`oldest-found-eth-block`-`latest-block`, we make not more than 20 calls
6. when all blocks which contain incoming/outgoing transfers for a
given address are found, we save these blocks to db and mark that
transfers from these blocks are still to be fetched
7. Then we select latest ~30 (the number can be adjusted) blocks from
these which were found and fetch transfers, this requires 3-4
requests per transfer.
8. we persist scanned range so that we know were to start next time
9. we dispatch an event which tells client that transactions are found
10. client fetches latest 20 transfers
- when user presses "fetch more" button we check if app's db contains next
20 transfers, if not we scan chain again and return transfers after
small fixes
2019-12-18 11:01:46 +00:00
func insertRange ( creator statementCreator , account common . Address , network uint64 , from * big . Int , to * big . Int ) ( err error ) {
insert , err := creator . Prepare ( "INSERT INTO blocks_ranges (network_id, address, blk_from, blk_to) VALUES (?, ?, ?, ?)" )
if err != nil {
return err
}
_ , err = insert . Exec ( network , account , ( * SQLBigInt ) ( from ) , ( * SQLBigInt ) ( to ) )
if err != nil {
return err
}
return
}
func updateOrInsertTransfers ( creator statementCreator , network uint64 , transfers [ ] Transfer ) error {
update , err := creator . Prepare ( ` UPDATE transfers
SET tx = ? , sender = ? , receipt = ? , timestamp = ? , loaded = 1
WHERE address = ? AND hash = ? ` )
if err != nil {
return err
}
insert , err := creator . Prepare ( ` INSERT OR IGNORE INTO transfers
( network_id , hash , blk_hash , blk_number , timestamp , address , tx , sender , receipt , log , type , loaded )
VALUES
( ? , ? , ? , ? , ? , ? , ? , ? , ? , ? , ? , 1 ) ` )
if err != nil {
return err
}
for _ , t := range transfers {
res , err := update . Exec ( & JSONBlob { t . Transaction } , t . From , & JSONBlob { t . Receipt } , t . Timestamp , t . Address , t . ID )
if err != nil {
return err
}
affected , err := res . RowsAffected ( )
if err != nil {
return err
}
if affected > 0 {
continue
}
_ , err = insert . Exec ( network , t . ID , t . BlockHash , ( * SQLBigInt ) ( t . BlockNumber ) , t . Timestamp , t . Address , & JSONBlob { t . Transaction } , t . From , & JSONBlob { t . Receipt } , & JSONBlob { t . Log } , t . Type )
if err != nil {
log . Error ( "can't save transfer" , "b-hash" , t . BlockHash , "b-n" , t . BlockNumber , "a" , t . Address , "h" , t . ID )
return err
}
}
return nil
}
//markBlocksAsLoaded(tx, address, db.network, blocks)
func markBlocksAsLoaded ( creator statementCreator , address common . Address , network uint64 , blocks [ ] * big . Int ) error {
update , err := creator . Prepare ( "UPDATE blocks SET loaded=? WHERE address=? AND blk_number=? AND network_id=?" )
if err != nil {
return err
}
for _ , block := range blocks {
_ , err := update . Exec ( true , address , ( * SQLBigInt ) ( block ) , network )
if err != nil {
return err
}
}
return nil
}