e3 - can't unwind (?)

[wmitsuda]

Using yesterday's main, commit 849e7a9592dfbaa5308644a60c290e1201e2dd34

I left my sepolia node a few days without syncing to the tip, yesterday I updated the main branch and left it fill out overnight, then got this issue, not sure if the node got corrupted or just some bug that blocks the unwind.

INFO[06-25|10:55:45.421] [EngineBlockDownloader] Processed        highest=6183602
INFO[06-25|10:55:45.421] Beginning downloaded blocks insertion
INFO[06-25|10:56:11.755] [insertHeadersAndBodies] Written blocks  progress=6160000 to=6183602
INFO[06-25|10:57:03.799] [p2p] GoodPeers                          eth68=3 eth67=1
INFO[06-25|10:57:03.801] [mem] memory stats                       alloc=2.5GB sys=12.8GB
INFO[06-25|10:57:03.915] [txpool] stat                            pending=0 baseFee=0 queued=30000 alloc=2.5GB sys=12.8GB
INFO[06-25|10:57:43.961] [sync] limited big jump                  from=6153963 amount=5000
WARN[06-25|10:57:44.192] [EngineBlockDownloader] Could not insert headers and bodies err="cannot unwind to block 6153963, lowest unwindable block is 18446744073709551615"
sync_locked:16893 skip update meta
INFO[06-25|11:00:03.789] [p2p] GoodPeers                          eth68=4 eth67=1
INFO[06-25|11:00:03.791] [mem] memory stats                       alloc=2.1GB sys=12.8GB
INFO[06-25|11:00:03.915] [txpool] stat                            pending=0 baseFee=0 queued=30000 alloc=2.1GB sys=12.8GB

[AskAlexSharov]

there was recently PR to unwind: https://github.com/ledgerwatch/erigon/pull/10903 try git pull and restart if it will not help - show logs

[banteg]

seeing the same on latest main (1797c66)

[INFO] [06-26|14:31:13.349] [sync] limited big jump                  from=20099999 amount=5000
[WARN] [06-26|14:31:14.169] Could not set forkchoice                 app=caplin stage=ForkChoice err="execution Client RPC failed to retrieve ForkChoiceUp date response, err: cannot unwind to block 20099999, lowest unwindable block is 18446744073709551615"
[WARN] [06-26|14:31:14.179] error executing clstage                  app=caplin stage=ForkChoice err="execution Client RPC failed to retrieve ForkChoiceUp date response, err: cannot unwind to block 20099999, lowest unwindable block is 18446744073709551615"

[Giulio2002]

mh, this is probably a bug that was there before though, I will investigate

[banteg]

tried again with 900f730f6c and the problem looks fixed. close after you get another confirmation from @wmitsuda.

[AskAlexSharov]

If just update didn’t help. Then next step is: update + “rm -rf chaindata caplin” cc: @banteg @wmitsuda

[wmitsuda]

I updated to yesterday’s main and after leaving it running overnight it seems fixed