kevko | morning | 05:49 |
---|---|---|
opendevreview | Michal Arbet proposed openstack/kolla-ansible stable/zed: Fix broken list concatenation in horizon role https://review.opendev.org/c/openstack/kolla-ansible/+/915018 | 06:04 |
kevko | ^^ mnasiadka, bbezak, frickler | 06:07 |
kevko | thanks | 06:07 |
mnasiadka | kevko: bbezak is off, just added +2 - just merge it (it's a backport, it doesn't matter you're the author) | 06:17 |
mnasiadka | kevko: but let's wait for zuul - let's hope it doesn't fail on 2.12 | 06:18 |
frickler | +1 to the let's wait on zuul, if it works fine, I can approve | 06:25 |
mnasiadka | ok then - time is ticking now for the kolla release - we have ~3 months at maximum, but would prefer to get there sooner than later | 07:44 |
mnasiadka | for kolla https://review.opendev.org/c/openstack/kolla/+/914744 needs to go in | 07:45 |
opendevreview | Michal Nasiadka proposed openstack/kolla master: prometheus: use inovex/prometheus-libvirt-exporter https://review.opendev.org/c/openstack/kolla/+/909507 | 07:53 |
opendevreview | Michal Nasiadka proposed openstack/kolla master: prometheus: use inovex/prometheus-libvirt-exporter https://review.opendev.org/c/openstack/kolla/+/909507 | 07:53 |
SvenKieske | would also be nice to get some reviews from cores on my list on the kolla whiteboard, should I crosspost here again? I asked for three weeks for reviews now, I think, and only mnasiadka took a look (thanks!) | 08:01 |
SvenKieske | mnasiadka: is that 2.15 in the toolbox commit a typo? I'm not sure I'm understanding the commit message | 08:03 |
mnasiadka | SvenKieske: no, rocky linux does not support 2.16, because it requires newer python | 08:04 |
SvenKieske | ah, well that would have been information for the commit message or something | 08:04 |
mnasiadka | we could think about installing newer python in rocky | 08:04 |
mnasiadka | maybe I'll try, but not today | 08:04 |
SvenKieske | if you do a commit message saying "bump ansible-core to 2.16" and your commit then bumps it to 2.15 for rpm, don't you expect questions? | 08:05 |
opendevreview | Michal Nasiadka proposed openstack/kolla master: kolla-toolbox: Bump ansible-core to 2.16 https://review.opendev.org/c/openstack/kolla/+/914744 | 08:06 |
SvenKieske | unfortunate that the inovex exporter is not backwards compatible :( but at least it exists :) | 08:07 |
SvenKieske | mhm, I guess it would not hurt to completely go the "pip install" route to provide a uniform environment even for the python base version? (regarding rocky linux). ansible-core had some surprisingly broken differences in behaviour in the last few releases imho | 08:13 |
SvenKieske | would be nice to align that in a next step, no? | 08:13 |
mnasiadka | Yeah, I'll follow up - but probably in a week or so - next week I'm on business trip in Washington DC | 08:14 |
SvenKieske | oh nice, that's quite a ride | 08:16 |
SvenKieske | I guess aligned versions will save us some headaches later on - hopefully. | 08:16 |
SvenKieske | mnasiadka: just tested in a fresh rocky container (rocky 9.3): default python version is 3.9 but you can just install "dnf install python3.11" but you have to call it via "python3.11" binary then | 08:39 |
mnasiadka | yeah, that's the plan, but needs some rework on more than one line in the Dockerfile | 08:39 |
mnasiadka | and I would prefer testing that in kolla-ansible extensively in case anything fails | 08:40 |
SvenKieske | sure | 08:41 |
SvenKieske | I don't use it myself, but maybe someone could review this swift fixes with regards to keystone auth? seems fairly straight forward: https://review.opendev.org/c/openstack/kolla-ansible/+/914965 | 08:51 |
SvenKieske | and of course the old topic of ovn-exporter: https://review.opendev.org/c/openstack/kolla-ansible/+/855498?usp=dashboard | 08:51 |
SvenKieske | mnasiadka: btw I started working on refactoring the prometheus scrap_config, is a little bit tricky, because we have differing variable and job names.. | 08:52 |
SvenKieske | would honestly be easier - I think - to template this with jsonnet instead of jinja https://jsonnet.org/, but I don't know if that would be okay, would introduce another tool again :-/ | 08:53 |
SvenKieske | I guess I try jinja first, will need some manual if checks I guess, but should be fine | 08:54 |
mnasiadka | we have every template in jinja2, that's ansible choice, let's not introduce something against the stream | 09:04 |
SvenKieske | yeah sure, jsonnet has some usage in the prometheus community to automate templating, whatever. I already thought I would get that answer :) | 09:06 |
SvenKieske | mhm, this is the second or third time CI says to me upgrade job is failing in "test-core-openstack.sh" but the log file does end in a successful test and just stops there? https://5a6297a260b050d49670-97f1bb270b356fde15ebdabc0499edaa.ssl.cf5.rackcdn.com/914967/1/check/kolla-ansible-rocky9-slurp-upgrade/f8c946e/primary/logs/ansible/test-core-openstack | 09:10 |
SvenKieske | is that a bug, a random CI failure? | 09:10 |
SvenKieske | I'll file a bug and collect some evidence :) | 09:11 |
SvenKieske | nice, now launchpad is throwing error pages at me as well | 09:12 |
mnasiadka | SvenKieske: https://5a6297a260b050d49670-97f1bb270b356fde15ebdabc0499edaa.ssl.cf5.rackcdn.com/914967/1/check/kolla-ansible-rocky9-slurp-upgrade/f8c946e/primary/logs/ansible/test-core-openstack-upgrade - this one | 09:15 |
mnasiadka | (it was test-core-openstack after upgrade.sh) :D | 09:15 |
mnasiadka | cirros kernel panic | 09:16 |
SvenKieske | but that is "test-core-openstack-upgrade" not "test-core-openstack" like zuul says? weird. | 09:17 |
SvenKieske | at least the launchpad loading failure is persistent if you try to report a new bug. reported that bug over on libera irc.. | 09:17 |
SvenKieske | they are working on it | 09:26 |
SvenKieske | thanks for the pointer, should've looked at the next log file I suppose. | 09:27 |
mnasiadka | would be probably good to fix the names a bit | 09:27 |
mnasiadka | but I would still probably think about moving most of the stuff to ansible | 09:27 |
SvenKieske | yes | 09:27 |
SvenKieske | mhm question regarding the kernel crash: are we using centrally managed (like versions selected etc) cirros images or do we do that on our own? | 09:28 |
SvenKieske | I can't remember having seen a definition of a cirros version in our code, at least, but that doesn't mean that much :D | 09:28 |
mnasiadka | we use cirros images released by cirros team on github and just pin to a version | 09:29 |
SvenKieske | ah we do | 09:29 |
SvenKieske | tools->init-runonce? | 09:30 |
SvenKieske | at least there is a version default set, but it's two years old? | 09:31 |
SvenKieske | doesn't seem to be set anywhere else | 09:32 |
opendevreview | Sven Kieske proposed openstack/kolla-ansible master: CI: bump cirros version https://review.opendev.org/c/openstack/kolla-ansible/+/915117 | 09:33 |
mnasiadka | well, 0.6.2 seems to have a newer kernel | 09:33 |
SvenKieske | a patch doesn't hurt ^^ let's see if it breaks | 09:34 |
SvenKieske | these are only very minor kernel bumps | 09:36 |
opendevreview | Michal Nasiadka proposed openstack/kolla master: Add deployment of fwaas https://review.opendev.org/c/openstack/kolla/+/914855 | 09:49 |
opendevreview | Michal Nasiadka proposed openstack/kolla master: horizon/neutron: Add FWaaS plugin https://review.opendev.org/c/openstack/kolla/+/914855 | 09:49 |
opendevreview | Sven Kieske proposed openstack/kolla-ansible master: CI: test s3api with keystone auth https://review.opendev.org/c/openstack/kolla-ansible/+/914967 | 10:03 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: opensearch: Add /api/status unauthenticated access https://review.opendev.org/c/openstack/kolla-ansible/+/915119 | 10:07 |
SvenKieske | thanks for the further feedback in ovn-exporter, some good questions raised, will investigate if we can get away with only deploying to nb/sb ovn hosts. | 10:11 |
opendevreview | Michal Nasiadka proposed openstack/kolla-ansible master: WIP: OpenSearch TLS https://review.opendev.org/c/openstack/kolla-ansible/+/915121 | 10:37 |
opendevreview | Jake Hutchinson proposed openstack/kayobe master: Register baremetal compute nodes in Ironic https://review.opendev.org/c/openstack/kayobe/+/909671 | 12:52 |
opendevreview | Roman KrĨek proposed openstack/kolla-ansible master: Performance: use filters for servicel dicts https://review.opendev.org/c/openstack/kolla-ansible/+/914997 | 14:38 |
opendevreview | Sven Kieske proposed openstack/kolla-ansible master: Add ovn-exporter https://review.opendev.org/c/openstack/kolla-ansible/+/855498 | 15:23 |
opendevreview | Winicius Allan Bezerra da Silva proposed openstack/kolla-ansible master: Change the default ceph_nova_user in globals https://review.opendev.org/c/openstack/kolla-ansible/+/915169 | 16:51 |
greatgatsby_ | hoping someone can offer some troubleshooting advise. Upgraded from yoga to zed, VMs no longer getting IPs via DHCP. I've figured out the qdhcp namespace and IP on the controller, and the tuntap and qbr on the compute. Did a packet capture on tap interface, can see the DHCP request going out with no response. `ip -s link show dev tap<foo>` only has RX bytes, no TX bytes. | 18:29 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!