ChainSafe / ChainSafe/gossamer
devnet node fails to initiate next BABE epoch due key missing from authority set
- Dominant language
- Go
- Stars
- 454
- Forks
- 144
- PR merge metrics
- No merged PRs in 30d
Description
## Task summary
at around block 10000 on the devnet, the Alice node displayed the following error:
```
2021-11-27T15:54:55Z INFO epoch 110 complete, upcoming epoch: 111 babe.go:L442 pkg=babe
2021-11-27T15:54:55Z DBUG initiating epoch 111 epoch.go:L18 pkg=babe
2021-11-27T15:54:57Z EROR failed to initiate epoch 111: key not in BABE authority data babe.go:L351 pkg=babe
2021-11-27T15:54:57Z CRIT block authoring error: key not in BABE authority data babe.go:L337 pkg=babe
```
this caused block production to cease for this node (other nodes were ok.)
this should not happen, as the gssmr runtime is set to have the same authority set every epoch (see https://github.com/noot/substrate/commit/f462f52b22ab5b14c08d9895cdc3cba3ca7ca507)
determine why this is happening and fix it. additionally, if BABE fails to initiate an epoch, it should retry at the next epoch instead of completely returning.
Contributor guide
Assessment
This issue has not been assessed yet.