| opendevreview | Merged openstack/python-octaviaclient master: Add extra validation for HSTS related arguments https://review.opendev.org/c/openstack/python-octaviaclient/+/892533 | 06:55 |
|---|---|---|
| opendevreview | Gregory Thiemonge proposed openstack/octavia-tempest-plugin master: Add additional subnets in LoadBalancerBaseTest class https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/832199 | 09:26 |
| opendevreview | Gregory Thiemonge proposed openstack/octavia-tempest-plugin master: Add member test with multiple subnets on same network https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/815313 | 09:26 |
| opendevreview | Gregory Thiemonge proposed openstack/octavia-tempest-plugin master: Add tests for additional VIPs https://review.opendev.org/c/openstack/octavia-tempest-plugin/+/664462 | 09:26 |
| opendevreview | Tom Weininger proposed openstack/octavia master: Integrate mypy type checker https://review.opendev.org/c/openstack/octavia/+/879749 | 10:06 |
| opendevreview | Tom Weininger proposed openstack/python-octaviaclient master: Fix inconsistency in validation error message for HSTS https://review.opendev.org/c/openstack/python-octaviaclient/+/892902 | 10:13 |
| opendevreview | Tom Weininger proposed openstack/octavia master: Integrate mypy type checker https://review.opendev.org/c/openstack/octavia/+/879749 | 10:15 |
| *** gthiemon1e is now known as gthiemonge | 12:20 | |
| gthiemonge | this is the 2nd time I see that on c9s: https://zuul.opendev.org/t/openstack/build/0dff0c4b69c94d14ae1fe2642f997b77/log/controller/logs/octavia-amphora_log.txt#25703-25724 | 13:28 |
| gthiemonge | haproxy killing itself? | 13:28 |
| opendevreview | Tom Weininger proposed openstack/python-octaviaclient master: Fix validation of HSTS attributes https://review.opendev.org/c/openstack/python-octaviaclient/+/892972 | 14:46 |
| opendevreview | Tom Weininger proposed openstack/python-octaviaclient master: Fix validation of HSTS attributes https://review.opendev.org/c/openstack/python-octaviaclient/+/892972 | 15:04 |
| opendevreview | Merged openstack/python-octaviaclient master: Fix inconsistency in validation error message for HSTS https://review.opendev.org/c/openstack/python-octaviaclient/+/892902 | 15:34 |
| opendevreview | Merged openstack/python-octaviaclient master: Fix validation of HSTS attributes https://review.opendev.org/c/openstack/python-octaviaclient/+/892972 | 15:36 |
| johnsom | Yeah, that thread kill in C9s is troubling.... | 15:41 |
| gthiemonge | johnsom: I'm wondering if it could be a memory allocation issue. there are a lot of frontends (I can check if we reach the 1M maxconn) + a couple of keepalived instances | 15:44 |
| johnsom | Yeah, I looked at that. It seemed it had 130MB free at the time | 15:45 |
| johnsom | Which is inconclusive really | 15:45 |
| gthiemonge | yeah but a reload allocates a lot of memory in a new worker | 15:46 |
| johnsom | Yeah, we usually see a memory error for haproxy though, not stuck threads.... I guess I didn't check the swap usage, I probably should have done that. | 15:46 |
| johnsom | This puzzles me a bit too: updt_fd_polling+0xc8/0x176 is it polling on a UDP socket for something? I don't think we have that enabled | 15:48 |
| johnsom | There is a lot of swap left available. So, I don't think that would stick a thread. | 15:50 |
| johnsom | lol, a few hits: https://github.com/haproxy/haproxy/issues?q=is%3Aissue+stuck%3D1 | 15:52 |
| gthiemonge | updt_fd_polling (for update) not UDP | 15:54 |
| gthiemonge | (for update?) | 15:55 |
| johnsom | Yeah, could be | 15:55 |
| gthiemonge | yeah a lot of people are reporting the same error message | 15:55 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!