michaelscheung

Results 17 comments of michaelscheung

I think I found the culprit. This seems to happen in state sync blocks. For instance, the failing example I gave for is in block , which has the following...

I've synchronized some blocks since the commit 30df8a9, but I did not resynchronize the entire blockchain. Retrieving logs with just the block number crashes with the same error: ``` user@UbuntuAmd:/mnt/raid_array/erigon/erigon-release$...

Please forgive the delay, I haven't been able to test it on any new blocks because of the following error: ``` [INFO] [08-20|00:52:06.818] [6/16 Execution] Blocks execution from=32009919 to=32103600 [WARN]...

> @michaelscheung no worries we had a state root hash error, and this might be related to that. @enriavil1 Ah I see, I think it is indeed related as I...

> mainnet? or it's some private mining devnet? @AskAlexSharov Confirming that I've encountered the same error on polygon during synchronization and the same error when trying to unwind. Erigon version...

Narrowed it down to integration_stage_trie: ``` user@UbuntuAmd:/mnt/raid_array/erigon/erigon-release$ ./build/bin/integration stage_trie --unwind=100 --datadir="/mnt/md4/erigon/polygon/erigon/datadir" --chain="bor-mainnet" INFO[08-20|12:37:33.264] StageExec progress=32009769 INFO[08-20|12:37:33.264] StageTrie progress=32009919 INFO[08-20|12:37:33.266] [9/16 IntermediateHashes] Unwinding from=32009919 to=32009819 csbucket=AccountChangeSet INFO[08-20|12:37:33.270] [9/16 IntermediateHashes] Unwinding from=32009919...

> Have you ever fix it? @michaelscheung @fynn-z I'm currently attempting to re-synchronize a backup I took two months ago on an older version of Erigon. I'll let you know...

Nice detective work @fynn-z! Does this mean that the node will still need to be resynchronized since the data is already corrupted and can't be unwound? Given the very long...

@AskAlexSharov Below was the error I got when I tried to unwind state_stages (stage_headers works fine). I had tried unwinding anywhere from 10 to 10,000 blocks, same error. I tried...

I did unwind state_stages as well as all of the individual stages on the latest version of Erigon, but kept encountering the same error. Resetting the stage_trie sounds like it...