fix startup after BN exited between head and finalized blocks updates (…

…#5617)

When the BN exits after writing new `head` to database, but before
completing the `updateFinalizedBlocks` call, the database is slightly
inconsistent due to the partial write. We currently fail to start up
after that. Fix that by catching up on partial `updateFinalizedBlocks`
tasks on start up, and add a test for this edge case.

AllTests-mainnet.md

@@ -638,8 +638,9 @@ OK: 3/3 Fail: 0/3 Skip: 0/3
 + init with gaps [Preset: mainnet]                                                           OK
 + orphaned epoch block [Preset: mainnet]                                                     OK
 + prune heads on finalization [Preset: mainnet]                                              OK
++ shutdown during finalization [Preset: mainnet]                                             OK
 ```
-OK: 3/3 Fail: 0/3 Skip: 0/3
+OK: 4/4 Fail: 0/4 Skip: 0/4
 ## createValidatorFiles()
 ```diff
 + Add keystore files [LOCAL]                                                                 OK
@@ -720,4 +721,4 @@ OK: 2/2 Fail: 0/2 Skip: 0/2
 OK: 9/9 Fail: 0/9 Skip: 0/9
 
 ---TOTAL---
-OK: 409/414 Fail: 0/414 Skip: 5/414
+OK: 410/415 Fail: 0/415 Skip: 5/415

beacon_chain/consensus_object_pools/blockchain_dag.nim

@@ -1008,7 +1008,6 @@ proc init*(T: type ChainDAGRef, cfg: RuntimeConfig, db: BeaconChainDB,
     # state - the tail is implicitly finalized, and if we have a finalized block
     # table, that provides another hint
     finalizedSlot = db.finalizedBlocks.high.get(tail.slot)
-    newFinalized: seq[BlockId]
     cache: StateCache
     foundHeadState = false
     headBlocks: seq[BlockRef]
@@ -1128,28 +1127,37 @@ proc init*(T: type ChainDAGRef, cfg: RuntimeConfig, db: BeaconChainDB,
 
   doAssert dag.finalizedHead.blck != nil,
     "The finalized head should exist at the slot"
-  doAssert dag.finalizedHead.blck.parent == nil,
-    "...but that's the last BlockRef with a parent"
 
   block: # Top up finalized blocks
     if db.finalizedBlocks.high.isNone or
         db.finalizedBlocks.high.get() < dag.finalizedHead.blck.slot:
+      # Versions prior to 1.7.0 did not store finalized blocks in the
+      # database, and / or the application might have crashed between the head
+      # and finalized blocks updates.
       info "Loading finalized blocks",
         finHigh = db.finalizedBlocks.high,
         finalizedHead = shortLog(dag.finalizedHead)
 
-      for blck in db.getAncestorSummaries(dag.finalizedHead.blck.root):
+      var
+        newFinalized: seq[BlockId]
+        tmp = dag.finalizedHead.blck
+      while tmp.parent != nil:
+        newFinalized.add(tmp.bid)
+        let p = tmp.parent
+        tmp.parent = nil
+        tmp = p
+
+      for blck in db.getAncestorSummaries(tmp.root):
         if db.finalizedBlocks.high.isSome and
             blck.summary.slot <= db.finalizedBlocks.high.get:
           break
 
-        # Versions prior to 1.7.0 did not store finalized blocks in the
-        # database, and / or the application might have crashed between the head
-        # and finalized blocks updates.
         newFinalized.add(BlockId(slot: blck.summary.slot, root: blck.root))
 
-  let finalizedBlocksTick = Moment.now()
-  db.updateFinalizedBlocks(newFinalized)
+      db.updateFinalizedBlocks(newFinalized)
+
+  doAssert dag.finalizedHead.blck.parent == nil,
+    "The finalized head is the last BlockRef with a parent"
 
   block:
     let finalized = db.finalizedBlocks.get(db.finalizedBlocks.high.get()).expect(

tests/test_blockchain_dag.nim

@@ -640,6 +640,43 @@ suite "chain DAG finalization tests" & preset():
       dag2.finalizedHead.slot == dag.finalizedHead.slot
       getStateRoot(dag2.headState) == getStateRoot(dag.headState)
 
+  test "shutdown during finalization" & preset():
+    var testPassed: bool
+
+    # Configure a hook that is called during finalization while the
+    # database has been partially written, to test behaviour if the
+    # beacon node is exited while the database is inconsistent.
+    proc onHeadChanged(data: HeadChangeInfoObject) =
+      if data.epoch_transition:
+        # Check test assumption: Head block was written before this callback
+        let headBlock = dag.db.getHeadBlock().expect("Valid DB")
+        doAssert headBlock == data.block_root, "Head was written before CB"
+
+        # Check test assumption: New finalized blocks were not written yet
+        let
+          stateFinalizedSlot =
+            dag.headState.getStateField(finalized_checkpoint).epoch.start_slot
+          dbFinalizedSlot =
+            dag.db.finalizedBlocks.high.expect("Valid DB")
+        doAssert stateFinalizedSlot > dbFinalizedSlot, "Finalized not written"
+
+        # If the beacon node were to exit _now_, this is what the DB looks like.
+        # Validate that we can initialize a new DAG from this database.
+        let validatorMonitor2 = newClone(ValidatorMonitor.init())
+        discard ChainDAGRef.init(
+          defaultRuntimeConfig, db, validatorMonitor2, {})
+        testPassed = true
+    dag.setHeadCb(onHeadChanged)
+
+    for blck in makeTestBlocks(
+        dag.headState, cache, int(SLOTS_PER_EPOCH * 4), attested = true):
+      let added = dag.addHeadBlock(verifier, blck.phase0Data, nilPhase0Callback)
+      check: added.isOk
+      dag.updateHead(added[], quarantine, [])
+      dag.pruneAtFinalization()
+
+    check testPassed
+
 suite "Old database versions" & preset():
   setup:
     let