Neutron Havana gate failure
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 Hi all, multiple backports are failing due to test_schedule_poll_with_disabled_agent in Havana [1][2]. There is a bug [3] for the failure that is resolved as Won't Fix. This failure blocks merges in the branch, so I propose to skip the test in the gate to allow remaining bits to enter the last release tomorrow. I've added that skip to my patch [2]. Any objections? This email can also be considered as an exception request for my patch [2]. The issue that it fixes was observed in production. The issue is that some ports may be missed from updates (not VLAN tagged) if they temporarily disappear from the integration bridge. This results in no network connectivity for instances bound to those ports. [1]: https://review.openstack.org/#/c/114851/ [2]: https://review.openstack.org/115984 [3]: https://bugs.launchpad.net/neutron/+bug/1299683 -----BEGIN PGP SIGNATURE----- Version: GnuPG/MacGPG2 v2.0.22 (Darwin) iQEcBAEBCgAGBQJUGUfYAAoJEC5aWaUY1u573+MIAM5/Z/ffvfPKqLbGoUye/joh VOJe6qrHlHOcQbroy8yQ3Yl1wiiaSZr7ABpmolcGvlS+R56q5dKdREnsGAPPj9u+ JPX2FiOd+RpVuKkvTUzUv5yYI3rJAdAnqXFKX11go47InRrswWr/ZNgU8JdyxhuJ 9NcSCICHpC64ZzyGsHBhqHeCOlYSpQ3fYEHLPkqgkNvmt2vmnuTZJwddOGxJ1x1L 2+CF1dXDN3eiE2ceAxB2Yv3R7PHyMPk+lcX3TzKrGu+ogsX3uUYlukkruk1nlzKR CuVtT8q/rmyGzSKohK3was1JwJQKnDdspidUrz2P3KLhFPU7x8ueqOcc1gsib+0= =aSBo -----END PGP SIGNATURE-----
This email can also be considered as an exception request for my patch [2]. The issue that it fixes was observed in production. The issue is that some ports may be missed from updates (not VLAN tagged) if they temporarily disappear from the integration bridge. This results in no network connectivity for instances bound to those ports.
IMHO this exception is justfied and included already in Icehouse so shouldn't bring regression. Disabling test is suboptimal but given comments in the bug, I don't other way around it. I've +2 leaving approval to Neutron PTL. Cheers, Alan
On Wed, Sep 17, 2014 at 10:11 AM, Alan Pevec <apevec@gmail.com> wrote:
This email can also be considered as an exception request for my patch [2]. The issue that it fixes was observed in production. The issue is that some ports may be missed from updates (not VLAN tagged) if they temporarily disappear from the integration bridge. This results in no network connectivity for instances bound to those ports.
IMHO this exception is justfied and included already in Icehouse so shouldn't bring regression. Disabling test is suboptimal but given comments in the bug, I don't other way around it. I've +2 leaving approval to Neutron PTL.
I agree with Alan's comments, and I've pushed this into the merge queue now. Thanks for pushing this one out Ihar! Kyle
Cheers, Alan
participants (3)
-
Alan Pevec
-
Ihar Hrachyshka
-
Kyle Mestery