Commit Graph

170 Commits

Author SHA1 Message Date
Agnish Ghosh 1120fc295a
add parallel recons to recons from gossip and background reconstruction 2024-11-17 23:14:50 +07:00
Agnish Ghosh 520788195d
cleanup, use nim taskpools 2024-11-15 17:32:37 +07:00
Agnish Ghosh ed5cf4cd8d
refactored reconstruction 2024-11-14 14:14:02 +07:00
Agnish Ghosh f17f5ecf94
fix a few more things 2024-11-14 11:14:40 +07:00
Agnish Ghosh 60889d5c5d
delay reconstruction 2024-11-14 00:27:03 +07:00
Agnish Ghosh b6cf733f91
clean up critical path 2024-11-13 18:33:32 +07:00
Agnish Ghosh e05cb51ffe
reconstruction only happens in background 2024-11-13 18:33:06 +07:00
Agnish Ghosh 8cc53a0428
make quarantine and db coherent 2024-11-13 15:30:54 +07:00
Agnish Ghosh 0f659a0c1f
revert to reconstructing from gossip first 2024-11-12 15:18:12 +07:00
Agnish Ghosh 68f979098d
enqueue block first, before reconstruction 2024-11-12 13:56:25 +07:00
Agnish Ghosh fae0079f94
fix database persistence issue 2024-11-12 11:18:34 +07:00
Agnish Ghosh fc8f3e0ea7
persist non finalized block data columns 2024-10-14 02:04:39 +05:30
Agnish Ghosh 1c1ff21bf9
other fixes 2024-10-14 01:20:51 +05:30
Agnish Ghosh 1cf1a62367
make sure quarantine persistent across all threads 2024-10-13 17:11:37 +05:30
Agnish Ghosh 1f3e2a94db
missing data column logic 2024-10-13 16:21:59 +05:30
Agnish Ghosh 6fc5fc79ed
have gather data columns in enqueue block 2024-10-13 13:48:51 +05:30
Agnish Ghosh 198ac66e2c
add more checks to ensure all custody columns are received 2024-10-13 04:22:23 +05:30
Agnish Ghosh 622cd0b087
slight bug fix 2024-10-13 01:52:37 +05:30
Agnish Ghosh 4a5995797f
fix 2024-10-12 19:20:36 +05:30
Agnish Ghosh dfcc89138f
add more specifics to reconstruction and block addition conditions 2024-10-12 16:42:56 +05:30
Agnish Ghosh 18fe511cc2
attempt to recover columns while backfilling 2024-10-12 03:41:23 +05:30
Agnish Ghosh 6ae1418e91
fix reconstruction from gossip logic 2024-10-12 01:48:18 +05:30
Agnish Ghosh 826a4f60f8
have 2 distinct functions for having enough and missing data columns 2024-10-11 21:42:28 +05:30
Agnish Ghosh ae46c85136 modify gatherDataColumns logic 2024-10-10 14:32:38 +05:30
Agnish Ghosh d456b99e37 fix 2024-10-10 05:03:21 +05:30
Agnish Ghosh a0a8ebfe22 slight fix in recon from gossip 2024-10-10 05:02:53 +05:30
Agnish Ghosh 1d80f7608d attempt reconstruction from gossip itself 2024-10-10 04:34:39 +05:30
Agnish Ghosh 4a006b5584 fix every place to make new custody column function work 2024-09-09 02:09:24 +05:30
Agnish Ghosh 9be615dff9
add: data column reconstruction and broadcast (#6481)
* save commit, decouples reconstruction and broadcasting

* save progress

* add: reconstruction event loop, previous reconstruction related cleanups
2024-08-08 17:44:55 +05:30
Agnish Ghosh b32205de7c
upstream peerdas alpha3 related spec changes + fix upstream related issues (#6468)
* reworked some of the das core specs, pr'd to check whether whether the conflicting type issue is centric to my machine or not

* bumped nim-blscurve to 9c6e80c6109133c0af3025654f5a8820282cff05, same as unstable

* bumped nim-eth2-scenarios, nim-nat-traversal at par with unstable, added more pathches, made peerdas devnet branch backward compatible, peerdas passing new ssz tests as per alpha3, disabled electra fixture tests, as branch hasn't been rebased for a while

* refactor test fixture files

* rm: serializeDataColumn

* refactor: took data columns extracted from blobs during block proposal to the heap

* disable blob broadcast in pd devnet

* fix addBlock in message router

* fix: data column iterator

* added debug checkpoints to check CI

* refactor if else conditions

* add: updated das core specs to alpha 3, and unit tests pass
2024-08-05 19:27:39 +05:30
Agnish Ghosh 85db9ca99c
regressive fix 2024-07-06 00:33:15 +05:30
Agnish Ghosh 152d276d78
added reconstruction logic 2024-07-03 21:50:52 +05:30
Agnish Ghosh 67fe8aca0b
disable blob activity (exp), improve gossip validation 2024-07-02 14:36:44 +05:30
Agnish Ghosh 5bf1e021a7
initiate data column quarantine 2024-06-28 14:53:08 +05:30
Agnish Ghosh 325bdfd4a2
support for enqueueing whichever is activated blob/data_column 2024-06-25 18:08:00 +05:30
Agnish Ghosh 791d2fb0d1
add: forward and backward syncing for data columns, broadcasting data columns created from blobs, added dc support to sync_queue 2024-06-24 17:32:06 +05:30
Agnish Ghosh 46d07b140d
add: data column support in sync_protocol, sync_manager, request_manager, fix: gossipValidation rules 2024-06-18 19:01:56 +05:30
Agnish Ghosh aa2ce79770
fixed upto electra 2024-05-27 18:10:00 +05:30
tersec c1b9e82502
electra attestation gossip plumbing (#6287) 2024-05-14 19:01:26 +03:00
tersec 3f972a2ca5
add some support for electra aggregated attestations (#6283) 2024-05-14 06:12:35 +02:00
tersec d139c92df9
explicitly scope AttesterSlashing and IndexedAttestation types to phase0 (#6224) 2024-04-21 05:49:11 +00:00
tersec 603c83522e
explicitly refer to phase0.{Attestation,TrustedAttestation} rather than sans module name (#6214) 2024-04-17 20:44:29 +00:00
tersec e51c5ec783
add Electra blob support to block/blob quarantines, block processor, and request manager (#6201) 2024-04-11 09:31:39 +00:00
tersec 7fa32b7f02
add Electra to ConsensusFork enum (#6169)
* add Electra to ConsensusFork enum

* fix gnosis check
2024-04-03 16:43:43 +02:00
Etan Kissling 6f466894ab
answer `RequestManager` queries from disk if possible (#6109)
When restarting beacon node, orphaned blocks remain in the database but
on startup, only the canonical chain as selected by fork choice loads.
When a new block is discovered that builds on top of an orphaned block,
the orphaned block is re-downloaded using sync/request manager, despite
it already being present on disk. Such queries can be answered locally
to improve discovery speed of alternate forks.
2024-03-21 18:37:31 +01:00
tersec 134774e00d
ensure reason field logging consistently uses string type (#5878) 2024-02-10 03:50:31 +01:00
Etan Kissling b7026a683a
avoid marking blocks as unviable if `blobless` quarantine is full (#5858)
Full caches should not be used to mark blocks as unviable. The unviable
status is quite persistent and a block marked as such won't be processed
again once the cache empties. Problem originally introduced in #4808.
2024-02-07 13:38:20 +00:00
tersec 45b4b46041
use "reason" instead of "error"/"validatorError" to log gossip ignore/reject reasons (#5839) 2024-01-31 03:18:20 +00:00
Jacek Sieka 6328c77778
raises for gossip (#5808)
* raises for gossip

* fix light client
2024-01-22 17:34:54 +01:00
Etan Kissling e7794b6aec
log `delay` relative to deadline (#5769)
To better analyze logs across different `SECONDS_PER_SLOT`, log delays
for duties relative to their deadline instead of relative to slot start.
2024-01-17 15:52:19 +01:00