Created attachment 1338108 [details]
Description of problem:
Change https://review.gluster.org/#/c/18485/ posted for fixing the errors in hooks scripts reported with Shellcheck utility failed to run regressions successfully on CentOS machines. Following test case failed in consecutive runs:
However it ran successfully on my local system (Fedora 26). So requesting a CentOS machine to root cause the issue.
You should be able to SSH in as firstname.lastname@example.org. Please delete any temporary files after you're done.
Is this activity complete? I really expect loaners to be back in 1 week with 2 being maximum.
Actually I was trying to reproduce the issue locally on a CentOS 6 machine. I was unable to do so. https://review.gluster.org/#/c/18485/2/extras/hook-scripts/create/post/S10selinux-label-brick.sh has some background.
For now you can take back the machine.
So, when I was working on https://review.gluster.org/#/c/17681/ I had a conversation with Niels about this. It works because as long as we get the context applied to the top level folder, the restorecon will make it work on the rest.
Machine back in the pool.
(In reply to Nigel Babu from comment #4)
> So, when I was working on https://review.gluster.org/#/c/17681/ I had a
> conversation with Niels about this. It works because as long as we get the
> context applied to the top level folder,
As of now it is not even being applied to the top level directory via hook-scripts. That was my point.