| tkajinam | croelandt, the issue currently affects all repos which don't use upper constraints during doc build. As frickler a few keystone repos are also affected, but there might be a few more | 00:55 |
|---|---|---|
| tkajinam | croelandt, note that the glance repo uses u-c in doc build and that's why we don't see the same problem here. so adding u-c fixes inconsistency | 03:52 |
| tkajinam | s/here/there/ | 03:52 |
| opendevreview | Abhishek Kekane proposed openstack/glance-specs master: Add spec to bypass chunked transfer of data https://review.opendev.org/c/openstack/glance-specs/+/947423 | 06:13 |
| mrjoshi | abhishekk, rosmaita, dansmith, croelandt glance weekly meeting in 10 minutes at #openstack-meeting | 13:51 |
| croelandt | rosmaita: Hey! I've tried to address https://bugs.launchpad.net/glance/+bug/1799491 with two patches https://review.opendev.org/c/openstack/glance/+/932350 and https://review.opendev.org/c/openstack/glance/+/944381 , take a look if you got a minute :) | 14:49 |
| opendevreview | Merged openstack/glance-tempest-plugin master: Add stable/2025.1 job https://review.opendev.org/c/openstack/glance-tempest-plugin/+/947111 | 14:49 |
| rosmaita | croelandt: ack ... when the heck did i file all these bugs? | 14:50 |
| croelandt | rosmaita: 2018 :) | 14:52 |
| rosmaita | must have been a slow year | 14:52 |
| croelandt | 25 bugs by you on Glance | 14:53 |
| croelandt | gotta close them all! | 14:53 |
| rosmaita | and those are only the ones i reported, there are probably others that i added to the code | 14:53 |
| rosmaita | croelandt: left a question for you on https://review.opendev.org/c/openstack/glance/+/932350 | 15:01 |
| croelandt | yeah you're right, the patch is old | 15:06 |
| croelandt | is the date supposed to be the date of the creation of the tag? | 15:06 |
| opendevreview | Cyril Roelandt proposed openstack/glance master: Update CLI doc header for 29.0.0 https://review.opendev.org/c/openstack/glance/+/932350 | 15:08 |
| opendevreview | Cyril Roelandt proposed openstack/glance master: Update CLI doc header for 30.0.0 https://review.opendev.org/c/openstack/glance/+/932350 | 15:08 |
| rosmaita | croelandt: i don't think it needs to be the exact date of the tag, just needs to be "pretty close" | 15:18 |
| JayF | rosmaita or some other glance expert: We wanted a question answered in an ironic review if possible https://review.opendev.org/c/openstack/ironic/+/947115/1/ironic/common/glance_service/service_utils.py#209 | 15:19 |
| opendevreview | Merged openstack/glance master: Update CLI doc header for 30.0.0 https://review.opendev.org/c/openstack/glance/+/932350 | 15:38 |
| croelandt | rosmaita: ok so my script works :) | 15:43 |
| rosmaita | oh, i know, i tested it out! | 15:44 |
| croelandt | wow you don't trust my shell abilities | 15:47 |
| croelandt | you're damn right about that | 15:47 |
| croelandt | cannot write two lines of shell without shellcheck and I still get confused about POSIX shell vs bashisms | 15:48 |
| rosmaita | JayF: left a comment, you are correct | 15:53 |
| rosmaita | JayF: i noticed all that swift-tmp-url stuff ... you do check to make sure the image status is 'active' before actually consuming the image in that case, right? | 15:55 |
| rosmaita | at some point glance added the ability for an operator to 'deactivate' and image without deleting its data to block image use while investigating whether there was something bad in the image | 15:55 |
| JayF | When swift use is enabled, the consumption of the image might be time-delayed from the check/generation of the URL by $serverPostTime + $deployStepsBeforeImageWriteTime | 15:56 |
| JayF | which might be anywhere from 5 minutes to 30 minutes | 15:56 |
| JayF | but we can't do a late check because that's on a remote agent with no control plane access | 15:56 |
| rosmaita | jeez, i thought that was supposed to speed things up! | 15:57 |
| JayF | you do realize physical servers can take 20+ minutes to post these days? | 15:57 |
| JayF | that's most of what that wait is :) | 15:57 |
| JayF | I'm not even counting cases of deploy steps that could require a reboot; like a just-in-time firmware update | 15:58 |
| JayF | but in those cases ... swift temp urls are likely not awesome to use anyway b/c of timeouts | 15:58 |
| JayF | (and honestly; swift temp url is not the most common use case for us these days -- most folks have the conductor pull the image from glance then stream it from conductor onto the agent/target) | 15:58 |
| rosmaita | well, presumably for that use case, the user knows exactly what they are dealing with, image-wise | 15:58 |
| JayF | yeah, Ironic just has weird cases around time delay; there are not many other cases where "server create" -> "server goes ACTIVE" can take 1/2 hr+ | 15:59 |
| opendevreview | Abhishek Kekane proposed openstack/glance master: Refactor scrubber to remove use of eventlet https://review.opendev.org/c/openstack/glance/+/947637 | 20:27 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!