openstackgerrit | Merged openstack/networking-bgpvpn master: Drop nose dependencies https://review.openstack.org/592672 | 01:22 |
---|---|---|
*** tmorin has joined #openstack-net-bgpvpn | 07:13 | |
*** bfernando has joined #openstack-net-bgpvpn | 07:42 | |
doude | hi tmorin | 08:26 |
doude | about https://review.openstack.org/#/c/600731, why is it unstable? | 08:26 |
doude | for https://review.openstack.org/#/c/600957 I don't knwo what it does. I never write a playbook and don't know how OpenStack infra works... | 08:28 |
tmorin | hi doude | 11:51 |
tmorin | test_router_association_update is just a unit test | 11:51 |
tmorin | whether you run it locally or via openstack CI playbooks does not really matter | 11:52 |
tmorin | tox -epy27 ... | 11:52 |
tmorin | I don't know exactly why it is not stable (or I would fix it), but it looks like a concurrency issue | 11:52 |
tmorin | I suspect seomthing like if multiple parallel runs register/unregister the advertise_extra_routes, in an inconsistent way | 11:54 |
tmorin | and this parameters ends up being unrecognized even though the bgpvpn-routes-control extension (that defines it) is loaded | 11:55 |
tmorin | or perhaps the extension is unloaded | 11:55 |
doude | hmm interesting issue tmorin but not easy to tackle | 11:57 |
doude | so for the moment you prefer to keep it unstable to not block CI? | 11:58 |
tmorin | doude: yep, exactly | 13:00 |
tmorin | when it fails, we'll still see it (if we look at skipped tests), but it won't block merges | 13:01 |
doude | approved | 14:09 |
*** bfernando has quit IRC | 16:15 | |
*** tmorin has quit IRC | 16:18 | |
*** bfernando has joined #openstack-net-bgpvpn | 18:33 | |
*** bfernando has quit IRC | 19:43 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!