Red Hat Bugzilla – Bug 1303119
'pcs cluster cleanup' returns non-zero after couple of iterations
Last modified: 2017-10-09 15:48:13 EDT
Created attachment 1119455 [details]
pcs cluster report output
Description of problem:
I'm having a running cluster with no constraints and no resources.
I run in iteration create cluster (and constraints), start all resources, run
pcs cleanup, remove all resources. After several (usally tenths) of iteration
'cleanup' returns error.
Version-Release number of selected component (if applicable):
How reproducible: needs about 20 iterations to happen
Steps to Reproduce:
1. run several tenths of iterations of the regression test we use
Actual results: 'pcs resource cleanup --wait' returns non-zero
Expected results: 'pcs resource cleanup --wait' returns zero
I was not able to isolate the exact commands necessary but it is happening
always with our regression (python) tests. I was hoping that you might be able
to have a look in the logs I'm attaching to see whether you can find out why
this is happening.
Created attachment 1119456 [details]
pcs cluster cib
Created attachment 1119457 [details]
pcs cluster config
Some questions to help target what's going on:
Do your scripts run "pcs cleanup" without a specific resource name?
Do you have any timestamps from when the cleanup returned nonzero?
Do you know the exact value that was returned?
Do you know whether this occurs on other RHEL versions?
We have a fairly short time frame to get any fixes into 6.9. We will do a rebase to upstream 1.1.15 for 6.9, which fixes a number of bugs. If you can answer any of the questions in Comment 4, we can investigate further, otherwise we'll have to close this one.
Since RHEL 6 is beyond the "Production 2" phase and will only get critical bug fixes, I am closing this without a fix. However, there have been rebases since this report, so it may no longer be an issue.