On 8/7/20 4:56 PM, Christian Göttsche wrote:
Is this maybe related to #720096 and #831764 ?
logrotate calling the rsyslog-rotate wrapper twice (one time for
/var/log/syslog and another time for the rest (/var/log/mail.* etc.))
and the second time rsyslog is not (yet) ready to receive the SIGHUP ?
I created #720096, but I doubt that its a race condition in logrotate
in this case. As shown above I can reproduce the problem on the command
line (without logrotate) by running
systemctl kill -s HUP rsyslog.service
Not all hosts are affected, though. I haven't found a reliable way to
reproduce the problem yet.
Regards
Harri
