| opendevreview | Merged openstack/hacking master: Fix nova integration job https://review.opendev.org/c/openstack/hacking/+/874516 | 06:20 |
|---|---|---|
| opendevreview | Natz cheng proposed openstack/devstack master: remove unicode prefix from code https://review.opendev.org/c/openstack/devstack/+/896341 | 06:36 |
| frickler | ^^ why do people keep doing this? are they struggling just to do any contribution at all? I'm close to giving up on this, just to make it stop. https://review.opendev.org/c/openstack/devstack/+/854528 https://review.opendev.org/c/openstack/devstack/+/852318 https://review.opendev.org/c/openstack/devstack/+/890887 | 07:03 |
| frickler | kopecmartin: any progress on global_venv for centos? IMO it would be good if we could deprecate not using it for the next cycle, currently it seems lots of projects prefer to just disable it instead of fixing their testing | 07:08 |
| frickler | also, should we drop openeuler before we branch? | 07:08 |
| kopecmartin | frickler: unfortunately not yet, don't have any updates on that .. regarding openeuler, based on the latest comment in https://review.opendev.org/c/openstack/devstack/+/894796 it seems it'll be fixed | 07:11 |
| frickler | kopecmartin: oh, I didn't see that comment. weird timing, too. but o.k., then we can leave it as is for now | 07:16 |
| opendevreview | Yadnesh Kulkarni proposed openstack/devstack stable/xena: Use RDO official CloudSIG mirrors for C9S deployments https://review.opendev.org/c/openstack/devstack/+/896215 | 09:25 |
| opendevreview | Yadnesh Kulkarni proposed openstack/devstack stable/xena: Use RDO official CloudSIG mirrors for C9S deployments https://review.opendev.org/c/openstack/devstack/+/896215 | 09:26 |
| kopecmartin | frickler: i gave the global_venv bug almost 2 hours, but I absolutely don't understand it, there is no obvious pattern, the issue is somehow related to keystone, but how? I don't see the logic in stack.sh where we have 'if GLOBAL_VENV' and symlinking random service - i say random because there are more commands in /opt/stac/data/venv/bin than those that are symlinked and i don't see the pattern :/ | 09:50 |
| kopecmartin | also some services, e.g. lib/glance, have 'if GLOBAL_VENV' but some don't like e.g. the failing keystone o.O | 09:50 |
| kopecmartin | i tried adding 'sudo ln -sf /opt/stack/data/venv/bin/keystone-manage /usr/local/bin' in stack.sh, didn't help | 09:52 |
| kopecmartin | i'm starting to think, this isn't devstack's bug, it might be keystone's | 10:01 |
| frickler | kopecmartin: so as to why only some are symlinked: I only created symlinks for those that I saw causing failures, trying to keep that list minimal. maybe linking all or adding /opt/stac/data/venv/bin to PATH might be a better solution after all | 10:11 |
| kopecmartin | let me try that since i already have an env | 10:18 |
| kopecmartin | 'sudo ln -sf /opt/stack/data/venv/bin/* /usr/local/bin || true' .. let's see what happens :D | 10:20 |
| kopecmartin | nope, only that is not enough , hmm, keystone failed to open /opt/stack/data/venv/bin/keystone-wsgi-public .. cause it's not there, what creates that? | 10:29 |
| frickler | kopecmartin: hmm, in my devstack that file exists, same timestamp as keystone-manage | 10:33 |
| frickler | kopecmartin: https://opendev.org/openstack/keystone/src/branch/master/setup.cfg#L57-L64 seems like centos might place wsgi_scripts somewhere else? | 10:34 |
| kopecmartin | keystone-manage and keystone-status are there (/opt/stack/data/venv/bin/), only keystone-wsgi-public is missing | 10:35 |
| kopecmartin | oh, let me search the whole system | 10:35 |
| kopecmartin | keystone-wsgi-public doesn't exist anywhere o.O | 10:36 |
| kopecmartin | so maybe something that was supposed to create it failed, wasn't run etc | 10:37 |
| frickler | hmm, then try something like "pip install -v keystone", not sure what option you need in order to see in detail what pip does | 10:38 |
| frickler | also check the pip version in the venv | 10:38 |
| frickler | with "pip -vvv" is see something like: Installing keystone-wsgi-admin script to build/bdist.linux-x86_64/wheel/keystone-23.1.0.dev49.data/scripts | 10:45 |
| frickler | but no mention of the script getting installed later. only about removing it when doing "pip -U" | 10:46 |
| frickler | at least I can confirm that pip does indeed install that file on bookworm | 10:51 |
| * kopecmartin back from lunch | 11:38 | |
| kopecmartin | pip install -v keystone did install the file, i can see it now at /home/cloud-user/.local/bin/keystone-wsgi-public | 11:38 |
| frickler | kopecmartin: oh, may that be a special case for editable installs? I didn't try that yet | 11:57 |
| frickler | interesting, for editable install this does get mentioned explicitly: "Installing keystone-wsgi-public script to /opt/stack/data/venv/bin" | 12:06 |
| frickler | kopecmartin: is_suse indeed still needs to be dropped from the designate plugin, but that should not cause any failures, since "command not found" has the same effect as is_suse returning false | 12:51 |
| kopecmartin | frickler: right, i just wanted to prove the fact that most likely the plugin is outdated | 12:54 |
| opendevreview | Merged openstack/devstack master: Use OS_CLOUD in sample local.sh https://review.opendev.org/c/openstack/devstack/+/885764 | 13:31 |
| gmann | kopecmartin: do you want to +W on these tempest releases series? https://review.opendev.org/c/openstack/tempest/+/895984 https://review.opendev.org/c/openstack/tempest/+/895985/2 https://review.opendev.org/c/openstack/tempest/+/895986/2 | 17:11 |
| opendevreview | Artom Lifshitz proposed openstack/whitebox-tempest-plugin master: Use old value for service report interval https://review.opendev.org/c/openstack/whitebox-tempest-plugin/+/891612 | 18:39 |
| kopecmartin | gmann: absolutely | 18:59 |
| opendevreview | Merged openstack/tempest master: Add releasenote to tag the Tempest for 2023.2 release https://review.opendev.org/c/openstack/tempest/+/895984 | 20:04 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!