[Bugs] [Bug 1615003] Not getting logs if test case is time out on build server

bugzilla at redhat.com bugzilla at redhat.com
Sat Aug 11 18:20:13 UTC 2018


https://bugzilla.redhat.com/show_bug.cgi?id=1615003

Shyamsundar <srangana at redhat.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |srangana at redhat.com



--- Comment #2 from Shyamsundar <srangana at redhat.com> ---
This is because the test is calling cleanup twice in its exit bash traps.

- First, the test itself sets a trap to cleanup at
https://github.com/gluster/glusterfs/blob/master/tests/bugs/core/multiplex-limit-issue-151.t#L30

- There is an additional trap set to cleanup in include.rc,
https://github.com/gluster/glusterfs/blob/master/tests/include.rc#L719

The tar ball is generated in the cleanup routine, and also ensures that
on content in the tar balls is between 2 invocations. Thus, calling
cleanup twice will result in an empty tarball.

This can be seen running the test locally as,
`./tests/bugs/distribute/bug-1042725.t`

There are a few things in that test we need clarified,
1. why trap this:
https://github.com/gluster/glusterfs/blob/master/tests/bugs/core/multiplex-limit-issue-151.t#L29
2. Why trap cleanup, rather than invoke it at the end of the test as is
normal

Also, in the merged patch sets 2/4/6/7/8 (of [1]) I had added a cleanup at the
end (as I traced the failure of ./tests/bugs/distribute/bug-1042725.t to
incorrect cleanup by the previous test (or timeout in cleanup)). I did
not do the same in patch set 9.

So, I will post a patch that removes the traps set by this test (so we
get logs from this test), and hence add the manual cleanup at the end of
the test. (see bug#1615037 for the same)

Finally, I do not see an infra bug in this.

[1] https://review.gluster.org/c/glusterfs/+/20637

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are the assignee for the bug.


More information about the Bugs mailing list