|
== Attendees ==
Lars Kurth
Anthony Perard
Joao Martins
== OPEN ACTIONS ==
ACTION: Jim to investigate where such a change needs to be made and whether it is invasive (NOTE: Jim notes that this may take some time)
Jim: Sorry I haven't found time to make progress on this.
ACTION: Joao will go through documentation and let Lars know what to change
Joao: did not have time to do this
== Agenda ==
* CI Loop Report 1
I'd like to give you an update about the CI loop, what I've been doing,
the changes I've made, and the bugs.
Now, the list of tempest tests to skip are in nova.git tree (once
in Nova, he can update the list of skip tests.
Some new tests in tempest always fail, I've open bugs for them, and add
them to the list of skipped tests:
- test_volume_swap
- test_create_image_from_paused_server
There is another bug for which I apply a patch locally to tempest,
because sometime a test fail with "No IPv4 addresses found", the patch
workaround it, but might not be the solution.
I've updated the CirrOS image been used as a guest, the image have a
newer kernel, 4.8.0-42.45 from Ubuntu. That seems to help with suspend
failure. In anycase, there is also a patch for nova in the CI loop that
should print the guest console after a suspend failure.
I'll update this email thread from time to time.
Joao: looked at the Neutron issue, but nothing spring to mind. Suspend issue was solved
Joao: looks OpenStack related, and was going to reach out on a Neutron developer, but has not responded yet.
----
I did restarted the CI loop last week. It is about 70% success, so there
is still 30% of failure, mostly false.
There is one issue that I've notice that is often why a run fail, it is
something to do with attaching an extra volume.
For example, in
tempest.log (console.html):
Details: volume add66099-b0b8-43fd-ad92-5dc5850f1ff9 failed to reach in-use status (current available) within the required time (196 s).
libvirt/libxl-driver.log:
libxl: debug: libxl_aoutils.c:88:xswait_timeout_callback: backend /local/domain/0/backend/vbd/131/51744/state (hoping for state change to 2): xswait timeout (path=/local/domain/0/backend/vbd/131/51744/state
libxl: debug: libxl_event.c:867:devstate_callback: backend /local/domain/0/backend/vbd/131/51744/state wanted state 2 timed out
libxl: error: libxl_device.c:1006:device_backend_callback: unable to add device with path /local/domain/0/backend/vbd/131/51744
libxl: error: libxl.c:2014:device_addrm_aocomplete: unable to add vbd with id 51744
I don't know what's wrong, and I'm not sure where to look next. Any idee?
Anyway, there some nice graph about the success rates here:
Joao: never seen this issue before
Anthony: looked at the XML, which contains the configuration. Check out Nova CPU log to find it
ACTION (Joao): will look into this issue or try and find someone to do so
----
* Tempest Series in OSSTEST
Lars: still stuck on some of the infrastructure issues in the Test Lab which are currently with Credativ, before we can take that into service. There were still some issues to be resolved which need chasing.
ACTION (Lars): will raise this at the next call with Credativ.
_______________________________________________
Wg-openstack mailing list
| 
