proxysql - systemd - random failure to start after system crash

[kashak88]

proxysql 2.3.2
ubuntu 20.04

How to replicate:

1. Make sure proxysql is running and pid file exists (ex. /var/lib/proxysql/proxysql.pid and it has a process id inside, let's say 1010)
2. Imitate a "crash" by resetting a vm instead of doing a graceful shutdown (ex. in ESXI, right click - power - reset)
3. System restarts and systemd tries to start proxysql

What happens (as expected):
proxysql[753]: Process 1010 died: No such process; trying to remove PID file. (/var/lib/proxysql/proxysql.pid)
proxysql starts successfully

What can also happen at random:
proxysql[742]: Daemon already running on PID file 1010
systemd[1]: proxysql.service: Control process exited, code=exited, status=1/FAILURE
systemd[1]: proxysql.service: Failed with result 'exit-code'.
systemd[1]: Failed to start High Performance Advanced Proxy for MySQL.

TLDR; some other service starts as PID 1010 after crash and systemd fails to start proxysql due to lingering pid file.
This can likely be fixed with Restart=on-failure set as default? Or maybe allow us to specify pid file location in proxysql.conf so it could be kept in a volatile location like /var/run/proxysql or something?

Regards

[renecannao]

Hi @kashak88 .
PR #3707 from @smeinecke probably solves your issue.
The feature will be available in the next release.
Thanks