Comment 12 for bug 1818682

Every VM stop and start causes haproxy to be restarted. Most of the time it is less than 3 seconds, that haproxy is not running and sometimes it is up to 30 seconds like documented above.

I don't see any difference in the configuration of haproxy, therefore it does not make any sense to restart it. Why is it restarted? And what makes this restart sometimes delay for 30 seconds. I don't see any significant load or iowait on the network or controller node during that time.

We will continue our rolling upgrade to Queens and Rocky and pray, that the problem might go away.