A long weekend run of all the tp-libvirt virsh tests was mostly successful; however, the failures that existed caused ripple effects in other tests...
As a reminder - I start with the top of the virt-test/tp-libvirt tree's, add in changes I have in pull requests (I have 5 right now and other 1 or 2 that I have to clean up), then on my f20 laptop running the top of the libvirt tree I run: ./run -t libvirt --tests virsh --no virsh.dompmsuspend I'm avoiding dompmsuspend right now because there is a hang somewhere with sending the command through the Guest Agent which I haven't been able to fully diagnose yet. This results in 3336 tests being run with the following results: PASS => 3031 FAIL => 26 SKIP => 279 RATE => 99.15% Of the failures - I traced them to recently pushed PR's or they were existing/known issues. For the recently pushed PR's - I logged in the PR what the result of the test was in hopes someone fixes them. The current failures are: - 10 virsh.volume_application (PR#60) - 2 virsh.vol_resize (known failure, bz#1077068) - 3 virsh.undefine (PR#141) - 1 virsh.snapshot_create_as (known guest agent issue - investigating) - 9 virsh.schedinfo_qemu_posix (known failure with PR#145) - 1 virsh.domblkerror (PR#48) >From the failures listed above a number of other tests also failed because of "entrails" left behind on the failure. The entrails are xml changes which caused future tests to fail because of the presence of something unexpected. For this run - a failure in the 3rd undefine FAIL to restore the xml caused failures in 'snapshot_edit' and 'snapshot_create_as' since a snapshot was left in the guest xml. The failure in 'domblkerror' to create a backup xml and restore it caused all the 'dommemstat' and 'blockcopy' tests run afterwards to fail because there was no backup xml and a volume added to the domain was no longer present. Fortunately for me 'domblkerror' was one of the last tests to run; otherwise, there would have been more failures. Running those tests separately after the full run was done had expected results. John _______________________________________________ Virt-test-devel mailing list [email protected] https://www.redhat.com/mailman/listinfo/virt-test-devel
