-
Notifications
You must be signed in to change notification settings - Fork 218
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
Subnet knows about pending tx, but has stopped producing blocks #1115
Comments
Tx has a nonce of 23. Did your chain accepted other 22 txs from this sender? What is your chain config? Could you send startup logs? |
Yeah, the nonce isn't it. I tried with several different mechanisms and several different accounts, including the faucet, ferdy flip dapp, and finally cast directly. Startup logs for validator: https://gist.github.com/tactical-retreat/5692eca862f7156602d77e3b2135aa98 |
I have no chain config for this subnet, just an upgrade file (to enable the warp precompile and to force set the teleporter code at the right location). Genesis file is here: https://github.com/DeepWaterStudios/ff-subnet-public/blob/main/subnet-configuration/genesis-files/ferdy_testnet.json |
Could you run this and see what's reported there (assuming localhost;9650)?
|
FYI the RPC endpoint is https://testnet-rpc.ferdyflip.xyz/rpc if you want to test anything yourself. There's no explorer yet but I've been using https://devnet.routescan.io/ with that endpoint. |
Do you see anything in validator logs? Could you send a validator's log file (if they're all showing the same thing)? |
Validator startup logs were posted above, nothing exciting in them. Logs are very boring after that. If I send a tx:
I see:
Nothing after that. If I resend I get:
|
And to clarify what I mean by boring, just some C Chain output, mostly warp messages:
|
I added a chain config with logging set to debug and I now see these entries, which again confirm that it knows about pending tx. But nothing that indicates why there are no new blocks.
|
Actually, I see that my repo (https://github.com/tactical-retreat/ferdynet-subnet-evm) is one commit behind head, and the head commit has a lot of stuff related to gossip. I'll look into redeploying a new VM. |
OK I sync'd redeployed, and had the faucet send some tokens. On the RPC node I see:
On the validator I see:
|
All of your validators are running v0.6.0> and with same This shows a completely different chain: https://devnet.routescan.io/block/1. (last accepted block no is 8). |
Yes. I upgraded them yesterday to 1.11.2 + the subnet-evm compiled from head. They all have the same upgrade.json. I don't have the full logs from yesterday's run, sorry. I can probably restore to that if necessary. But I went ahead and recreated the subnet (FerdyNet2 now) to see if the issue was that my DB was borked, which is why you see a different chain now. The new chain is working fine, so apparently that's not it. The next thing I'm going to do is repeat the circumstances which got me into the previous problem, wipe the certificates, regenerate, re-add validators with 2000 weight. Will see if that makes this reproducable. |
OK, it seems like this is reproducible. Startup logs: https://gist.github.com/tactical-retreat/05b073ac017d1554f41c1d7de640f95d Reproduction instructions:
Tx will hang forever. |
3- Did you just delete the staking information or whole database? |
3 - just the staking information
5 - node id, bls key, pop. the stuff that core validation site requires that comes out of this command
6 - I used the exact same machines with the exact same db that they had when i stopped avalanchego |
We have recently introduced few changes that addresses the issue. Could you try with the recent versions of Subnet-EVM (0.6.4)and Avalanchego (1.11.5)? |
@tactical-retreat did this fix your issue? |
Describe the bug
I have 3 validator nodes that I registered prior to BLS signatures being required for registration. Since I want to use AWM, I had to wipe the staking info from the nodes, generate new node IDs, and re-register as validators.
To avoid having insufficient stake, I registered the nodes with 2000 stake instead of the default 20. Everything went pretty smoothly and the subnet came back up. I can fetch blocks. I can submit transactions.
But no new blocks are being produced, and there are no errors or clarifying info in the logs. The validators are definitely aware of the pending transaction. I logged into a validator and used
cast
to send some ether around:To Reproduce
Not sure how to reproduce this. or how to debug further.
I'm hoping there's some obvious thing I borked that can be fixed with a magical flag =(
The text was updated successfully, but these errors were encountered: