Comment 13 for bug 2067345

I checked this morning magnum-conductor status: it is failing with the exception mentioned in this issue, despite max_pool_size=0:

sqlalchemy.exc.TimeoutError: QueuePool limit of size 0 overflow 50 reached, connection timed out, timeout 30.00 (Background on this error at: https://sqlalche.me/e/14/3o7r)

which is a bit strange as 0 should mean no limit.... Looking at the MariaDB server log, there are a lot of

2024-07-07 23:59:43 544875 [Warning] Aborted connection 544875 to db: 'magnum' user: 'magnum' host: 'os-77023-admin' (Got timeout reading communication packets)

followed by:

Status information:
Current dir: /var/lib/mysql/
Running threads: 607 Cached threads: 9 Stack size: 299008

The number of thread looks very high for a test cloud with a very low activity...

On the Magnum server, counting the number of established connections the same way as previously, there are only 135 connections, a number quite low compared to what has been observed previously. There are 33 magnum-conductor processes running.

Michel