Open chrisguida opened 2 years ago
docker OOM kills aren't necessarily shown in dmesg
Are we able to find that in docker logs?
why does this depend on Start9Labs/embassy-os#1138?
It might not. I thought that was the "get resources under control" ticket. If it's not relevant we can remove it.
I'd rather not designate it as BLOCKED on 1138. Having the relation link is fine.
I had blocked on there because Aiden suggested that we wait until the system-level resource control is finished before revisiting this.
I believe this is the result of a race condition with the system clock. If this happens, it will cause a message to be spit into the logs that triggers the autoreindex logic.
The case above is not an autoreindex. Notice that only a couple hundred blocks were dropped. This is more likely to do with an ungraceful shutdown. The question is why the ungraceful shutdown occurred. The autoreindex/ntp bug is a separate issue
Does that other bug have a ticket or do I need to track it?
Just created one now: https://github.com/Start9Labs/embassy-os/issues/1193
Sometimes bitcoind just dies. This results in lost sync progress. I recently had this happen and it started resyncing 200 blocks back.
It is not being OOM killed, since there is no record of this in dmesg.
No idea what else it could be