opendevreview | Gregory Thiemonge proposed openstack/octavia stable/2023.1: DNM test Add octavia-grenade-slurp CI job https://review.opendev.org/c/openstack/octavia/+/878928 | 11:04 |
---|---|---|
opendevreview | Michal Arbet proposed openstack/octavia stable/zed: Fix octavia to accept [ipv6]:port https://review.opendev.org/c/openstack/octavia/+/878658 | 12:05 |
opendevreview | Michal Arbet proposed openstack/octavia stable/yoga: Fix octavia to accept [ipv6]:port https://review.opendev.org/c/openstack/octavia/+/878659 | 12:07 |
skraynev | gthiemonge: hello. I have the following case: during member batch update directly in batch_update_members method in controller_worker.py, I got traceback due to DB connection error. It happens on attempt to get list of members and as expected this method do not do retry on such exception. So my question is: does it make sense to add retry for oslo_db.DBConnectionError ? OR if I migrate to using jobboard it also should fix such faile | 12:18 |
skraynev | d scenario... (I have some doubts about second, because as I understand jobboard is for tasks, but here error happens before running task) | 12:18 |
gthiemonge | skraynev: Hi, jobboard does not resolve any DB connection errors, it will only resume some tasks if they are interrupted (like a controller worker is killed) | 12:27 |
gthiemonge | skraynev: what was the result of this DB connection error? in most of the case, octavia handles the exception and should set the resources to ERROR, we don't retry | 12:28 |
opendevreview | Gregory Thiemonge proposed openstack/octavia stable/2023.1: DNM test Add octavia-grenade-slurp CI job https://review.opendev.org/c/openstack/octavia/+/878928 | 12:34 |
skraynev | the issue looks like this part of the code is not covered by handling such traceback error. it happens here: https://github.com/openstack/octavia/blob/stable/yoga/octavia/controller/worker/v1/controller_worker.py#L492-L504 | 12:36 |
skraynev | so before running task, but in the batch_update_members function | 12:37 |
gthiemonge | skraynev: ok, if you have a DB error, it fails and the LB is stuck in PENDING_UPDATE, right? | 12:39 |
skraynev | gthiemonge: yes | 12:39 |
gthiemonge | skraynev: yeah I think we need to handle DB errors here, they are handled in the API, in the tasks/flows of the worker, but not in the code that prepares the flows | 12:40 |
gthiemonge | skraynev: could you open an issue on launchpad? https://bugs.launchpad.net/octavia | 12:41 |
skraynev | gthiemonge: sure, np, I will do. under the handling error: do you mean as retry? or just catch it and set "ERROR" state? | 12:41 |
gthiemonge | yeah I think we should catch the exception, and set the provisioning_status of the LB to ERROR | 12:42 |
opendevreview | Tom Weininger proposed openstack/octavia-tempest-plugin master: Updating Octavia tempest jobs on Ubuntu jammy (22.04) https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/861369 | 12:58 |
opendevreview | Gregory Thiemonge proposed openstack/octavia-tempest-plugin master: DNM/WIP Testing server output https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/877667 | 13:01 |
opendevreview | Gregory Thiemonge proposed openstack/octavia-tempest-plugin master: DNM: Capture cirros console logs https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/877832 | 13:01 |
skraynev | gthiemonge: done: https://bugs.launchpad.net/octavia/+bug/2015239 | 13:04 |
gthiemonge | skraynev: thanks | 13:07 |
skraynev | gthiemonge: no problem. You already helped me to understand, that jobboard does not help in this case | 13:08 |
opendevreview | Takashi Kajinami proposed openstack/octavia-dashboard master: Fix missing comma in example logger setting https://review.opendev.org/c/openstack/octavia-dashboard/+/879468 | 15:33 |
opendevreview | Merged openstack/octavia-dashboard master: Fix missing comma in example logger setting https://review.opendev.org/c/openstack/octavia-dashboard/+/879468 | 18:24 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!