Feature: Prevent loading an existing data directory for the wrong network (#2825)

* Prevent loading an existing data directory for the wrong network

* Fix and add more info

nimbus/common/chain_config_hash.nim

@@ -0,0 +1,84 @@
+# Nimbus
+# Copyright (c) 2021-2024 Status Research & Development GmbH
+# Licensed under either of
+#  * Apache License, version 2.0, ([LICENSE-APACHE](LICENSE-APACHE))
+#  * MIT license ([LICENSE-MIT](LICENSE-MIT))
+# at your option.
+# This file may not be copied, modified, or distributed except according to
+# those terms.
+
+{.push raises: [].}
+
+import
+  std/[typetraits, tables],
+  eth/common/base,
+  eth/common/times,
+  eth/common/hashes,
+  eth/common/addresses,
+  stew/endians2,
+  stint,
+  nimcrypto/sha2,
+  ./chain_config
+
+# ------------------------------------------------------------------------------
+# When the client doing initialization step, it will go through
+# complicated steps before the genesis hash is ready. See `CommonRef.init`.
+# If the genesis happen to exists in database belonging to other network,
+# it will replace the one in CommonRef cache.
+# That is the reason why using genesis header or genesis hash + ChainId is 
+# not a good solution to prevent loading existing data directory for 
+# the wrong network.
+# But the ChainConfig + raw Genesis hash will make the job done before
+# CommonRef creation.
+# ------------------------------------------------------------------------------
+
+# ------------------------------------------------------------------------------
+# Private helper functions
+# ------------------------------------------------------------------------------
+
+func update(ctx: var sha256, val: uint64 | UInt256) =
+  ctx.update(val.toBytesLE)
+
+func update(ctx: var sha256, val: ChainId | EthTime | NetworkId) =
+  ctx.update(distinctBase val)
+
+func update(ctx: var sha256, val: bool) =
+  ctx.update([val.byte])
+
+func update(ctx: var sha256, val: Hash32 | Bytes8 | Bytes32 | Address) =
+  ctx.update(val.data)
+
+func update[T](ctx: var sha256, val: Opt[T]) =
+  if val.isSome:
+    ctx.update(val.get)
+
+func update[K, V](ctx: var sha256, val: Table[K, V]) =
+  mixin update
+  for k, v in val:
+    ctx.update(k)
+    ctx.update(v)
+
+func update[T: object](ctx: var sha256, val: T) =
+  for f in fields(val):
+    ctx.update(f)
+
+func update[T: ref](ctx: var sha256, val: T) =
+  for f in fields(val[]):
+    ctx.update(f)
+
+# ------------------------------------------------------------------------------
+# Public functions
+# ------------------------------------------------------------------------------
+
+func calcHash*(networkId: NetworkId, conf: ChainConfig, genesis: Genesis): Hash32 =
+  var ctx: sha256
+  ctx.init()
+  ctx.update(networkId)
+  ctx.update(conf)
+  if genesis.isNil.not:
+    ctx.update(genesis)
+  ctx.finish(result.data)
+  ctx.clear()
+
+func calcHash*(networkId: NetworkId, params: NetworkParams): Hash32 =
+  calcHash(networkId, params.config, params.genesis)

nimbus/db/storage_types.nim

@@ -21,7 +21,7 @@ type
     canonicalHeadHash = 4
     slotHashToSlot = 5
     contractHash = 6
-    transitionStatus = 7
+    dataDirId = 7
     safeHash = 8
     finalizedHash = 9
     beaconState = 10
@@ -59,6 +59,10 @@ func canonicalHeadHashKey*(): DbKey {.inline.} =
   result.data[0] = byte ord(canonicalHeadHash)
   result.dataEndPos = 1
 
+func dataDirIdKey*(): DbKey {.inline.} =
+  result.data[0] = byte ord(dataDirId)
+  result.dataEndPos = 1
+  
 func slotHashToSlotKey*(h: openArray[byte]): DbKey {.inline.} =
   doAssert(h.len == 32)
   result.data[0] = byte ord(slotHashToSlot)

nimbus/nimbus_execution_client.nim

@@ -17,6 +17,7 @@ import
   metrics,
   metrics/chronicles_support,
   kzg4844/kzg,
+  stew/byteutils,
   ./rpc,
   ./version,
   ./constants,
@@ -24,7 +25,9 @@ import
   ./nimbus_import,
   ./core/eip4844,
   ./db/core_db/persistent,
-  ./sync/handlers
+  ./db/storage_types,
+  ./sync/handlers,
+  ./common/chain_config_hash
 
 from beacon_chain/nimbus_binary_common import setupFileLimits
 
@@ -153,6 +156,24 @@ proc setupMetrics(nimbus: NimbusNode, conf: NimbusConf) =
     nimbus.metricsServer = res.get
     waitFor nimbus.metricsServer.start()
 
+proc preventLoadingDataDirForTheWrongNetwork(db: CoreDbRef; conf: NimbusConf) =
+  let
+    kvt = db.ctx.getKvt()
+    calculatedId = calcHash(conf.networkId, conf.networkParams)
+    dataDirIdBytes = kvt.get(dataDirIdKey().toOpenArray).valueOr:
+      # an empty database
+      info "Writing data dir ID", ID=calculatedId
+      kvt.put(dataDirIdKey().toOpenArray, calculatedId.data).isOkOr:
+        fatal "Cannot write data dir ID", ID=calculatedId
+        quit(QuitFailure)
+      return
+
+  if calculatedId.data != dataDirIdBytes:
+    fatal "Data dir already initialized with other network configuration",
+      get=dataDirIdBytes.toHex,
+      expected=calculatedId
+    quit(QuitFailure)
+
 proc run(nimbus: NimbusNode, conf: NimbusConf) =
   ## logging
   setLogLevel(conf.logLevel)
@@ -192,6 +213,7 @@ proc run(nimbus: NimbusNode, conf: NimbusConf) =
         string conf.dataDir,
         conf.dbOptions(noKeyCache = conf.cmd == NimbusCmd.`import`))
 
+  preventLoadingDataDirForTheWrongNetwork(coreDB, conf)
   setupMetrics(nimbus, conf)
 
   let com = CommonRef.new(