mizu.eth

@AskAlexSharov > 1. Yes, erigon must not panic in this case, must return error. And likely it’s a bug. > > 2. Erigon can download Only snapshots from whitelist (old...

I had the same problem (Could not validate block err="Not implemented [...]) with fb0f811f19c41d2f005fc6173e4af1e6f01f6319. Reverting to v2022.09.03 fixed it.

I tried using #5510's commit but I ended up just getting a bunch of "[NewPayload] Previously known bad block hash=0x88e729e7bda2a8fd56b14ee32db574f2b2fb56670e3ef5b3bbf9f823e426d7be" in a never ending loop, stuck at block 15612204.

I don't know how to do that. I tried just restarting Erigon but it was still stuck with the "Previously known bad block" message. On the other hand, reverting to...

(oh, I see you gave instructions on how to rewind above, but it wasn't needed with 2022.09.03)

> Yes, you need to unwind so erigon doesn't count them as bad blocks anymore maybe restarting erigon will work but I don't remember well. Anyhow unwinding ~10 blocks should...

Thank you, @AskAlexSharov. > 2. yes, we can do it in 90% of cases, but not in 100% That would probably still help! At least that way I could prevent...

> * show your logs There's nothing special in the logs. Here's the full log: ``` [INFO] [09-22|09:23:46.305] Build info git_branch=HEAD git_tag=v2021.10.03-1902-gfb0f811f1-dirty git_commit=fb0f811f19c41d2f005fc6173e4af1e6f01f6319 [INFO] [09-22|09:23:46.305] Starting Erigon on Ethereum mainnet......

> add --pprof flag, when you see high mem usage - do go tool pprof -inuse_space -png http://127.0.0.1:6060/debug/pprof/heap > mem.png I can try this but the memory usage isn't unusually...

To reiterate: I did restrict the container's memory to 7GB for this test, otherwise I would have to wait potentially days to be able to reproduce the OOM kill on...