,StuartOn Tue, Jan 19, 2021 at 6:20 PM Strahil Nikolov <hunter86_bg@yahoo.com> wrote:I have just built a test cluster (centOS 8.3) for testing DRBD and it works quite fine.Actually I followed my notes from https://forums.centos.org/viewtopic.php?t=65539 with the exception of point
Hello all,
I think I found the problem.
On the fenced node after a restart of the cluster stack , I saw the following:
controld(dlm)[13025]: ERROR: Uncontrolled lockspace exists, system must reboot.
Executing suicide fencing
I was so focused on the DC logs, so I missed it.
I guess with HALVM
Your e-mail ended up in my spam mailbox.
Yes, you can do the following:
1. Increase token + consensus timeouts (check the man for the proper ratio)
2. Always set the node/cluster in maintenance and stop the cluster stack before
patching.
Best Regards,
Strahil Nikolov В понеделник, 28