| *** pmannidi is now known as pmannidi|Away | 00:03 | |
| opendevreview | Michał Dulko proposed openstack/kuryr-tempest-plugin master: Limit number of pod requests https://review.opendev.org/c/openstack/kuryr-tempest-plugin/+/830384 | 11:42 |
|---|---|---|
| dulek | Interesting fact, a lot of CI failures are due to logs being too long. It's a bug in subunit, not much we can do besides making sure we're not producing so many logs. :/ | 13:17 |
| dulek | I mean tempest lgos. | 13:17 |
| gryf | I thought about that - we're hammering k8s api a lot during waiting for the resources to be in ready state, perhaps it would be beneficial for implementing something like we have in kuryr code - like the backoff mechanism? | 13:29 |
| gryf | I know, that we might have wait significant time longer to test finish, but at least it will finish. wdyt dulek? | 13:31 |
| dulek | Hm… So my thinking at the moment is that I' | 13:31 |
| dulek | I'd consider supressing logs from K8s API client. | 13:31 |
| dulek | The problem is that this actually really helps with debugging. | 13:31 |
| gryf | yeah. | 13:31 |
| gryf | so maybe we need to set some sleep, before hammering k8s api? like a minute or five so? | 13:32 |
| dulek | You can see above one idea to improve this stuff. I've tried removing the logs it'll suppress and it seems like I got the file reduced from 12 to 6 MBs. | 13:32 |
| dulek | At the moment the golder standard in kuryr-tempest-plugin seems to be just 5 seconds. | 13:32 |
| dulek | A backoff would probably be best and most flexible. | 13:32 |
| dulek | Like exponential. | 13:32 |
| gryf | right. | 13:32 |
| dulek | Okay, let's see if my patch improves anything (peaking into Zuul it seems to) and if that won't be enough, I'll look into implementing proper exponentail backoff. | 13:35 |
| dulek | But first - there are still tests failing. :/ | 13:35 |
| gryf | I'm not sure what is causing it, but there is usually one or two tests failing, and usually on OVS gates. | 13:38 |
| opendevreview | Merged openstack/kuryr-kubernetes master: Fix prepoulation of ports pool definition https://review.opendev.org/c/openstack/kuryr-kubernetes/+/825173 | 13:43 |
| opendevreview | Michał Dulko proposed openstack/kuryr-tempest-plugin master: Limit number of pod requests https://review.opendev.org/c/openstack/kuryr-tempest-plugin/+/830384 | 14:22 |
| opendevreview | Michał Dulko proposed openstack/kuryr-tempest-plugin master: Improve test_egress_np_to_service_wo_selectors https://review.opendev.org/c/openstack/kuryr-tempest-plugin/+/830435 | 14:33 |
| dulek | gryf: A little improvement to a test you seem to have written. ^ I hope it'll make it more stable. | 14:34 |
| gryf | looking. | 14:37 |
| dulek | Aaand there's a new repo certs issue in the CI, just as I tried to test my patches. :) | 14:47 |
| opendevreview | Michał Dulko proposed openstack/kuryr-tempest-plugin master: Temporarily disable stable jobs https://review.opendev.org/c/openstack/kuryr-tempest-plugin/+/830503 | 17:00 |
| *** tbarron is now known as Guest252 | 19:40 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!