Just figured I'd share...  Over the last few months I periodically run
every single 'tp-libvirt' virsh test against recent upstream changes
(libvirt, virt-test, and tp-libvirt) mostly as a way to ensure recent
changes don't cause regressions. During this time - I've found a few
libvirt upstream issues and a few test issues prior to release
checkpoints - so all in all a good use of CPU cycles. A run such as this
takes a full day to complete - it's usually a weekend type run.

My most recent run (changes as of Fri 4/25 @ 10PM) had the following
results:

TOTAL: 3174
PASS : 2889
FAIL :   17
SKIP :  268
PCT  : 99.4% (PASS/(TOTAL-SKIP))

I ran using "./run -t libvirt --test virsh --no virsh.dompmsuspend". I
chose to not include 'dompmsuspend' since I believe there's an upstream
change that has caused the test to just wait forever for something that
won't happen. I haven't had the cycles to dig into the condition, but it
was running at some time in the recent past and I do remember seeing
changes go by upstream in this space - I just wasn't paying close
attention to them and have to find them again.

Of the 17 failures

 - 4 are fixed by a patch pushed today (and validated)
 - 2 are "known" failures w/ a bz.
 - 2 that I know are directly related to a recent upstream fix

I believe the rest have differences due to some upstream change in
behaviour that's not reflected in the test. More than likely a fix to
some problem made after the test was created - it's just a matter of
time to investigate and determine where the change was made in order to
make the correct change in the test to reflect the change.

My environment does have 6 'tp-libvirt' changes that are waiting for
review and merge.

John

_______________________________________________
Virt-test-devel mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/virt-test-devel

Reply via email to