Unexpected "Error restarting mysql" events

Question

We are running multiple Auroa MySQL Databases (5.7.mysql_aurora.2.10.1) with one read and one write node each and for the past couple of days I've been receiving error-events from the read-nodes that a restart of Mysql was failing, e.g.:

```
Error restarting mysql: java.io.IOException: No such process
Error restarting mysql: java.nio.file.FileSystemException: /proc/11147/stat: No such process
```

As far as I can tell the availability of the database nodes is not impaird and they also don't seem to be under particuarly heavy load and also should have plenty of RAM left. I was also able to restart the nodes myself manually, but these messages keep occuring every couple of hours. Is there something else I can do to investigate this issue?

Answer

This is most likely your Aurora Replica needing to restart for its maintenance. Check for metric AuroraReplicaLag if there have been some lags detected.

https://aws.amazon.com/premiumsupport/knowledge-center/aurora-read-replica-restart/

Answer

Thank you for your answer. I've checked the replication lag, however it is only about 19ms on my reader instances, so I don't think this is the reason. Additionally I've now started seeing these events ("Error restarting mysql: java.io.IOException: No such process") to start happening on the writer instances as well every couple of hours.

Answer

I talked with the AWS support and apparently this is an internal event which should not be "customer visible" and can be ignored:

> Although these events may indicate an error, they are meant for internal monitoring, and can be safely ignored.
> Hence, we are working on a fix for removing this messaging from Aurora logging. Until then, these messages will emitted from the logs and should be ignored. The Aurora MySQL service is otherwise fully operational after a reboot with this log messaging.

Answer

We are seeing the same problem on a single read/writer Aurora instance with no replicas. First occurence of this issue was on 2022-01-17 running 2.10.1.

We have upgraded meanwhile to 2.10.2 and this "Error restarting mysql: java.io.IOException: No such process" event still keeps re-appearing after upgrade / instance restart.

Unexpected "Error restarting mysql" events

Relevant content