systemd.service's TimeoutStopSec attribute is set to infinityshould be still 0

[andsel]

From systemd version 229 TimeoutStopSec switched from accepting 0 with the meaning turn off to infinity, however looking in changelog https://github.com/systemd/systemd/blob/a8c9824d2a98116cbd6e5689355894158eb73a57/NEWS#L5266-L5270 the 0 value is still accepted for backward compatibility. Some distributions like centos-7, could have the old version and setting it to infinity simply doesn't work as expected.

[juan-domenech]

+1 on this one.
Confirmed that Logstash 7.10 under Centos-7 Systemd shuts down immediately after stop command, leaving the service as failed and might cause data loss.

[root@testbox ~]# systemctl stop logstash
[root@testbox ~]# systemctl status logstash
● logstash.service - logstash
   Loaded: loaded (/etc/systemd/system/logstash.service; disabled; vendor preset: disabled)
   Active: failed (Result: signal) since Tue 2021-01-05 14:29:47 UTC; 1s ago
  Process: 347844 ExecStart=/usr/share/logstash/bin/logstash --path.settings /etc/logstash --log.level trace (code=killed, signal=KILL)
 Main PID: 347844 (code=killed, signal=KILL)
Jan 05 14:29:06 testbox systemd[1]: Started logstash.
Jan 05 14:29:47 testbox systemd[1]: Stopping logstash...
Jan 05 14:29:47 testbox systemd[1]: logstash.service stop-sigterm timed out. Killing.
Jan 05 14:29:47 testbox systemd[1]: logstash.service still around after SIGKILL. Ignoring.
Jan 05 14:29:47 testbox systemd[1]: logstash.service: main process exited, code=killed, status=9/KILL
Jan 05 14:29:47 testbox systemd[1]: Stopped logstash.
Jan 05 14:29:47 testbox systemd[1]: Unit logstash.service entered failed state.
Jan 05 14:29:47 testbox systemd[1]: logstash.service failed.

[juan-domenech]

As workaround I'm commenting out #TimeoutStopSec=infinity in /etc/systemd/system/logstash.service to use default settings.
This gives time to Logstash to shutdown gracefully.