Is this a bug report or feature request?
Deviation from expected behavior:
I think some time ago it was implemented that monitors were chosen letter consecutive indexes. While it looks like it does not really work.
This morning I had a little outage and out of 3 monitors: mon-a, mon-b, mon-e - the latter mon-e was unavailable. So after 10 minutes deadline mon-f was created. So I now run in a cluster with a, b, f.
What happens when I run out of alphabet?
Expected behavior:
I think letters should be assigned consecutively, if possible. In the case above it should be possible.
How to reproduce it (minimal and precise):
Make a monitor unavailable several times, and wait while it fails-over.
File(s) to submit:
cluster.yaml, if necessaryCrashing pod(s) logs, if necessary
To get logs, use kubectl -n <namespace> logs <pod name>
When pasting logs, always surround them with backticks or use the insert code button from the Github UI.
Read Github documentation if you need help.
Environment:
uname -a):rook version inside of a Rook Pod): rook: v1.3.5ceph -v): ceph version 14.2.9 (581f22da52345dba46ee232b73b990f06029a2a0) nautilus (stable)kubectl version):ceph health in the Rook Ceph toolbox):@zerkms When the single-letter alphabet is exhausted, a double-letter name will be used, then triple-letter, etc. The only time the letters increase is when there is a mon failover. What do you see is not working? Rook can't use the same mon name after a failover.
@travisn I think I seen it in some other issue report several years ago: that in the deployment a, b, e, on failover c would be used (to replace dead e), as it's available now. Or am I confusing it with something else?
@zerkms Rook has always added the next mon in alphabetical increasing order. There was a bug a few months back that caused rook to get confused about the index of the mon last used, but that was fixed.
Okay, so if exhausting is not causing any troubles - I don't mind. Thanks! :-)