-
Notifications
You must be signed in to change notification settings - Fork 20.1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Unhandled trie error in Trie.Prove
with experimental pbss
#26293
Comments
Starting a fresh sync this time with debug logs, in hopes of catching this error. |
Aside from |
|
Now that I'm watching for it, it doesn't occur again. Of course. Is this more likely to be an issue that just needs a specific state change to happen on chain to occur, and so it's just a matter of waiting; or more likely to be an issue that could be "forced" to reoccur by just syncing from scratch a bunch of times? |
It feels like some trie nodes are corrupted. Maybe you can run |
That particular DB is no longer available, I started over with a fresh DB and debug logs. To reproduce this, is it more likely that just waiting will, or that syncing from scratch repeatedly until issue is encountered will? |
I don't know, I think you can just try to play around with it and save the db if it happens again. |
Btw, I just did another round of snap sync and it can generate snapshot correctly. |
Unable to repro so far. I'll keep trying. |
for what it's worth, I also have these errors, tried to clean the DB multiple times but after the DB gets to ~500GB it start giving me these messages and never fully syncs. |
Are you also using this experimental PBSS branch? What do you mean "tried to clean the DB multiple times"?
Can you also share some logs? |
I've synced two of these pretty much once a day for the last week, and I cannot reproduce this now. If it's still in the code, it's some real corner case. |
@yorickdowne thanks for your time! I also fixed some issues in state history construction. But I guess it's irrelevant with your original issue, which should only occur when Since you are heavily testing/playing around it, I will let you know if I have the next stable version. |
@yorickdowne You may want to try out the latest version in #26274 Fixes some issues in trie history construction. Please sync from scratch if you want to have a try by any chance. |
@rjl493456442 Thanks! Will give it a whirl. |
It appears to have gotten "stuck" during sync. Moved the pivot and stopped making progress after:
It's now 16 hours later - no state download progress and no further receipts, just another
I have this DB in that state - what can I run for you? |
Well. Can you |
It's the other way around it turns out, exec has to be before the target: Attached. |
Thanks for the info. I don't find anything relevant with PBSS. Perhaps it's relevant with your network environment? Or the setup with consensus client? Does it eventually recover and finish the sync? |
System information
Geth version:
1.11.0-unstable
CL client & version: Teku
v22.11.0
Commit hash :
eb57884596df203c40c3144ed51542bb7e9ef3e8
, PR #26274Observation
On a fresh sync on mainnet with
--trie.path-based=true
, Geth is following chain head after sync and prints errors aboutmissing trie node 746d39f09b4c5c1b26ce016fcc3468c88eddecb18b0897b6e031d49925b8d1ab
This is a test system. It's understood pbss is experimental and not production-ready.
Logs
Logs when this error first occurred. I don't have debug logs, alas. This is not quite 2 hours after initial sync finished.
The text was updated successfully, but these errors were encountered: