| opendevreview | Ghanshyam proposed openstack/devstack master: WIP: testing sync mode https://review.opendev.org/c/openstack/devstack/+/811245 | 00:29 |
|---|---|---|
| frickler | oh, wow, that issue seems to be a regression in apache2 for ubuntu. at least that's the only thing that I can find changed. and c8 doesn't seem to be affected | 03:28 |
| opendevreview | Tushar Trambak Gite proposed openstack/os-testr master: Switch testing to Xena testing runtime https://review.opendev.org/c/openstack/os-testr/+/811293 | 03:38 |
| frickler | https://bugs.launchpad.net/ubuntu/+source/apache2/2.4.41-4ubuntu3.5 | 03:39 |
| opendevreview | Tushar Trambak Gite proposed openstack/os-testr master: Switch testing to Yoga testing runtime https://review.opendev.org/c/openstack/os-testr/+/811293 | 03:45 |
| opendevreview | Tushar Trambak Gite proposed openstack/openstack-health master: Switch testing to Yoga testing runtime https://review.opendev.org/c/openstack/openstack-health/+/811294 | 03:46 |
| frickler | o.k., I can confirm that an older, working devstack setup breaks in the same way after upgrading apache2 from 2.4.41-4ubuntu3.4 to 2.4.41-4ubuntu3.5 | 03:54 |
| opendevreview | Tushar Trambak Gite proposed openstack/os-performance-tools master: Switch testing to Yoga testing runtime https://review.opendev.org/c/openstack/os-performance-tools/+/811296 | 04:03 |
| opendevreview | Tushar Trambak Gite proposed openstack/stackviz master: Switch testing to Yoga testing runtime https://review.opendev.org/c/openstack/stackviz/+/811297 | 04:07 |
| opendevreview | Tushar Trambak Gite proposed openstack/tempest-stress master: Switch testing to Yoga testing runtime https://review.opendev.org/c/openstack/tempest-stress/+/811298 | 04:10 |
| opendevreview | Tushar Trambak Gite proposed openstack/devstack-tools master: Switch testing to Yoga testing runtime https://review.opendev.org/c/openstack/devstack-tools/+/811299 | 04:14 |
| opendevreview | Merged openstack/devstack-tools master: Switch testing to Yoga testing runtime https://review.opendev.org/c/openstack/devstack-tools/+/811299 | 04:40 |
| opendevreview | Merged openstack/tempest-stress master: Switch testing to Yoga testing runtime https://review.opendev.org/c/openstack/tempest-stress/+/811298 | 04:40 |
| frickler | seems to be an upstream issue really https://bugs.launchpad.net/ubuntu/+source/apache2/+bug/1945274 | 05:33 |
| ianw | frickler: bah, should have read your messages :) i just confirmed your findings! | 05:34 |
| ianw | https://git.launchpad.net/ubuntu/+source/apache2/tree/debian/patches/CVE-2021-36160.patch?h=applied/ubuntu/focal-security seems to be the patch | 05:35 |
| frickler | ianw: yes, I just tested with a revert of that patch | 05:35 |
| ianw | oh, i see you built without it too :) | 05:35 |
| frickler | I also pinged folks in #ubuntu-server and -security, so this hopefully will get fixed fast | 05:37 |
| frickler | I don't see an easy local workaround for us | 05:37 |
| frickler | although it is interesting that only the placement app seems to be affected and others like keystone are not. impact probably depends on the framework being used | 05:38 |
| ianw | u_path_info = ap_strchr(url + sizeof(UWSGI_SCHEME) + 2, '/'); | 05:40 |
| ianw | should that be 3 maybe, to strip ://? | 05:40 |
| ianw | i guess sizeof includes the null, so that "sizeof(UWSGI_SCHEME)" is the "uwsgi:" bit, and the +2 is for "//", and then it's saying "find the next /" | 05:45 |
| frickler | ianw: UWSGI_SCHEME is just "uwsgi", so that seems to make sense. will test with a modified patch | 05:51 |
| ianw | frickler: i think that bit might be right, perhaps ap_unescape_url is involved | 05:53 |
| ianw | frickler: ohh, i wonder if this has to do with it using a filesystem socket | 05:58 |
| ianw | $ ls /var/run/uwsgi/ | 05:59 |
| ianw | cinder-wsgi.socket keystone-wsgi-public.socket nova-api-wsgi.socket placement-api.socket | 05:59 |
| opendevreview | Dr. Jens Harbott proposed openstack/devstack master: DNM: test uwsgi fix https://review.opendev.org/c/openstack/devstack/+/811303 | 06:04 |
| frickler | ianw: ^^ that works locally for me. i.e. strip the final / in the proxypass definition | 06:04 |
| ianw | frickler: i agree. 104.130.219.63 is held and has that applied | 06:06 |
| ianw | (and restored to current apache) | 06:07 |
| ianw | /usr/local/bin/nova-status --debug --config-file /etc/nova/nova.conf upgrade check is happy | 06:08 |
| frickler | so is this actually a bug in our config that's uncovered by the apache CVE? | 06:17 |
| ianw | frickler: yeah, not 100% sure ... is there documentation on the| in the proxypass? | 06:23 |
| frickler | see https://httpd.apache.org/docs/trunk/mod/mod_proxy.html#proxypass - "If the first argument ends with a trailing /, the second argument should also end with a trailing /, and vice versa. Otherwise, the resulting requests to the backend may miss some needed slashes and do not deliver the expected results." | 06:24 |
| frickler | (thanks to amurray for pointing to that) | 06:25 |
| frickler | "miss some needed slashes" has now turned into "miss the last character" but in general that confirms the error on our side I'd say | 06:25 |
| ianw | frickler: interesting, does that mean https://uwsgi-docs.readthedocs.io/en/latest/Apache.html is wrong | 06:33 |
| ianw | ohh, no i see | 06:33 |
| ianw | dropped a comment on your change, seems we can either ensure ${url} has or does not have a trailing slash, and make sure ...${name} follows. i'm not sure which is better? | 06:38 |
| frickler | ianw: good point, it didn't actually occur to me that one could also just add a /, I think that's the better option likely as it should be easier to enforce | 06:51 |
| frickler | it will also be necessary to push this into oldest stable first in order to appease grenade | 06:51 |
| frickler | feel free to take over if you want, I'll be busy downstream for a couple of hours now | 06:52 |
| frickler | also someone may want to check this unrelated fedora failure https://zuul.opendev.org/t/openstack/build/4899915e09554652b05bb01d7a626a24 | 06:53 |
| ianw | frickler: unfortunately i'm also running up against my time limits here :) having identified this, it's probably good to loop some people like fungi in who might know if there's some corner case or other with either approach | 06:53 |
| opendevreview | Rajat Dhasmana proposed openstack/devstack master: Remove cinder from service names https://review.opendev.org/c/openstack/devstack/+/811310 | 07:13 |
| *** ianw is now known as ianw_pto | 07:21 | |
| *** jpena|off is now known as jpena | 07:31 | |
| kopecmartin | gmann: what's with the syntax error? unknown project openstack/horizon - https://review.opendev.org/c/openstack/tempest/+/810998 | 08:23 |
| tosky | kopecmartin: the error of the StorPool CI? Maybe a local configuration issue | 09:03 |
| kopecmartin | tosky: yeah, therefore I'm hesitating with merging the review , although, logically it's totally unrelated to the error | 09:16 |
| opendevreview | Soniya Murlidhar Vyas proposed openstack/tempest master: This patch creates the network v2.0 floating-ips-port-forwarding client. https://review.opendev.org/c/openstack/tempest/+/810015 | 09:29 |
| kopecmartin | gmann: https://review.opendev.org/c/openstack/tempest/+/808909 | 09:43 |
| tosky | kopecmartin: but it's an issue of their CI, not something anyone of us can fix AFAIK | 10:10 |
| opendevreview | Ian Wienand proposed openstack/devstack stable/xena: uwsgi: append url with a slash https://review.opendev.org/c/openstack/devstack/+/811389 | 10:12 |
| lyarwood | ianw: looks like that broke other endpoints https://zuul.opendev.org/t/openstack/build/22c9d99d0cb341cbb889be5e100816c9 /identity for keystone now returns a 404 | 10:42 |
| *** bhagyashris_ is now known as bhagyashris|rover | 11:16 | |
| frickler | o.k., I guess that settles the decision towards not using trailing slashes, lest we want to do the extra redirects that yoctozepto mentioned https://paste.opendev.org/show/809644/ | 11:20 |
| *** jpena is now known as jpena|lunch | 11:24 | |
| yoctozepto | frickler: I only mentioned them if we really need them | 11:27 |
| yoctozepto | otherwise let's skip it | 11:27 |
| *** dviroel|out is now known as dviroel | 11:27 | |
| frickler | yoctozepto: well for devstack itself we are fine, not sure if some plugin might be using write_uwsgi_config with an url that has a trailing slash | 11:30 |
| yoctozepto | frickler: I guess we can leave that out | 11:32 |
| frickler | well at least what I found on codesearch also looks fine | 11:32 |
| yoctozepto | ok | 11:33 |
| opendevreview | Dr. Jens Harbott proposed openstack/devstack stable/victoria: Fix uwsgi config https://review.opendev.org/c/openstack/devstack/+/811399 | 11:45 |
| *** jpena|lunch is now known as jpena | 12:17 | |
| amodi | artom: lyarwood any idea why these jobs are failing and now they're not even triggered https://review.opendev.org/c/openstack/whitebox-tempest-plugin/+/811193/1 https://review.opendev.org/c/openstack/whitebox-tempest-plugin/+/786979/35, the link for me is botched, i canot quite figure out the issue | 12:51 |
| artom | amodi, will look in about an hour | 13:02 |
| amodi | sure | 13:03 |
| rosmaita | amodi: all devstack-based jobs are currently hosed: http://lists.openstack.org/pipermail/openstack-discuss/2021-September/025101.html | 13:08 |
| amodi | since ysday right? | 13:10 |
| amodi | rosmaita: when're they going to be fixed | 13:10 |
| rosmaita | amodi: don't know, frickler and ianw_pto have some patches up ... you can read this channel's log to see the earlier discussion | 13:11 |
| rosmaita | looks like a change to the apache server to address a CVE did some bad things for us | 13:11 |
| rosmaita | amodi: conversation starts here: https://meetings.opendev.org/irclogs/%23openstack-qa/%23openstack-qa.2021-09-28.log.html#t2021-09-28T05:33:18 | 13:13 |
| amodi | rosmaita: will do thanks! | 13:16 |
| fungi | rosmaita: oh, is this fallout from ossa-2021-002 errata 1? | 13:58 |
| fungi | i'm in meetings all morning but trying to catch up now | 13:59 |
| rosmaita | fungi: i don't think so, i think this was an Apache CVE | 14:00 |
| rosmaita | fungi: CVE-2021-36160 | 14:01 |
| fungi | oh! | 14:09 |
| fungi | okay thanks | 14:09 |
| fungi | so it's new apache breaking things? | 14:09 |
| rosmaita | i think so, though it's not clear if it's because we're misusing something that they have now fixed, or they broke a legitimate use case | 14:15 |
| gmann | yeah it is not clear yet if it is breaking change or just our setup | 14:24 |
| clarkb | if you read the links that ianw and frickler posted it seems that the issue is our setup. We don't follow the rules set out by the docs | 14:51 |
| clarkb | your trailing /s have to match up between both sides of the proxy pass directive | 14:52 |
| clarkb | I suspect that one reason https://review.opendev.org/c/openstack/devstack/+/811303/1/lib/apache might not work is that sometimes $url has a / and sometimes it does not? I guess we can always strip the trailing / off of $url if it exists and then use fricklers change there | 14:53 |
| clarkb | that would be my suggestion anyway | 14:53 |
| opendevreview | Clark Boylan proposed openstack/devstack master: Fix apache ProxyPass directives with unbalanced trailing / https://review.opendev.org/c/openstack/devstack/+/811303 | 15:04 |
| clarkb | frickler: ^ I took the liberty of updating your DNM change. Hopefully that was ok (I figured since it was a -1 and DNM it isn't a big deal) And now I must meeting | 15:04 |
| gmann | yeah, I am not saying it is all they broke but as it used to work and doing it in security update seems odd to me. I think its both. | 15:16 |
| clarkb | the security update specifically had to do with exploits in proxy pass allowing you to act as an open proxy | 15:17 |
| clarkb | I'm sure they had to change all sorts of fun stuff in mod proxy | 15:17 |
| clarkb | oh I get it now the only failures were in grenade ebcause the old side hasn't updated. | 15:37 |
| gmann | yeah that is one part but it seems keytone broken on ianw_pto fix on xena. | 15:38 |
| gmann | trailing '/' | 15:38 |
| clarkb | Anyway I think the sanitization my update has made is worthwhile but that can happen in a followup if we do the rollup with the simpler fix | 15:38 |
| clarkb | gmann: right lets do no trailing /? | 15:39 |
| clarkb | is there a change for that on all of the branches that need it yet? | 15:39 |
| clarkb | I guess victoria has a ceph issue | 15:40 |
| gmann | yeah https://review.opendev.org/c/openstack/devstack/+/811399 | 15:40 |
| gmann | I think i have updated grenade on stable/xena setup let me do it on stable/xena and see | 15:41 |
| clarkb | Force merging the older branches is an option too if they fail on random things but grenade works. Then you can fix the random things later while unblocking the new release | 15:42 |
| clarkb | but before we do that we should haev a series of changes with depends on in the right order that we can start processing | 15:42 |
| clarkb | (and I don't see that existing anywhere yet) | 15:42 |
| gmann | yeah, it is not yet up. let me backport and see how things goes | 15:43 |
| clarkb | in this case you don't want to backport you want to forward port | 15:43 |
| clarkb | and you should use depends on across the branches so that grenade can be confirmed to work on each change | 15:44 |
| gmann | xena setup is merged in greande so we need to do on stable/xena too https://review.opendev.org/c/openstack/grenade/+/810994/1/roles/configure-grenade-branches/defaults/main.yaml | 15:44 |
| clarkb | My suggestion: On the old branch that needs the no trailing / fix push that change, forward port to the next change and depends-on the previous change. Repeat until you've made it up to master. | 15:46 |
| clarkb | That should fix things based on existing evidence. Force merge any changes that need extensive debugging for other issues. Approve the rest. Depends on Should handle it in the gate | 15:46 |
| clarkb | Then later you can decide if you need to followup with the url / trimming that my updated ps to the DNM master change does | 15:47 |
| *** jpena is now known as jpena|off | 16:31 | |
| opendevreview | Lee Yarwood proposed openstack/grenade master: WIP zuul: Add the neutron devstack plugin and neutron-trunk service https://review.opendev.org/c/openstack/grenade/+/811117 | 16:32 |
| opendevreview | Dr. Jens Harbott proposed openstack/devstack stable/victoria: Fix uwsgi config for trailing slashes https://review.opendev.org/c/openstack/devstack/+/811399 | 16:32 |
| frickler | clarkb: gmann: ^^ I dropped the failing jobs for now and will forward pick this change with deps | 16:33 |
| gmann | frickler: thanks.I think we need to do until stable/stein as stable/train still has the grenade jobs even it is in EM state | 16:34 |
| lyarwood | gmann: did the apache CVE land in Bionic? | 16:35 |
| frickler | gmann: grenade for victoria was passing, I'm assuming bionic isn't affected, possibly apache2 there was old enough | 16:35 |
| *** artom_ is now known as artom | 16:35 | |
| gmann | ok. | 16:35 |
| frickler | so victoria and newer should be sufficient | 16:35 |
| gmann | yeah, +1 | 16:35 |
| frickler | but I'll also make fixes for the older branches after that, they just aren't that urgent | 16:36 |
| gmann | frickler: or we can go with making grenade job on master->stable n-v and get master one merge soon by seeing release time and rc* fixes waiting ... | 16:38 |
| opendevreview | Dr. Jens Harbott proposed openstack/devstack stable/wallaby: Fix uwsgi config for trailing slashes https://review.opendev.org/c/openstack/devstack/+/811445 | 16:40 |
| frickler | I wouldn't feel good with non-working grenade jobs. this will also only be 4 patches, so shouldn't take long (famous last words, I know ;) | 16:42 |
| gmann | ok. | 16:43 |
| opendevreview | Dr. Jens Harbott proposed openstack/devstack stable/xena: Fix uwsgi config for trailing slashes https://review.opendev.org/c/openstack/devstack/+/811389 | 16:46 |
| clarkb | melwitt: I left a comment on https://review.opendev.org/c/openstack/devstack/+/811399 to help explain why you see the behavior you see there | 17:21 |
| *** slaweq_ is now known as slaweq | 17:22 | |
| clarkb | frickler: is there a change for master yet with the depends on chagne or do you just want to update https://review.opendev.org/c/openstack/devstack/+/811303 ? | 17:24 |
| clarkb | I've +2'd the changes for xena, wllaby, and victory and gave them gate blocker +2 priority status | 17:24 |
| gmann | yeah waiting for gate result on them. lgtm too | 17:24 |
| clarkb | hrm I think wallaby might have the ceph multistore problem. | 17:27 |
| opendevreview | Lee Yarwood proposed openstack/grenade master: zuul: Add the neutron devstack plugin and enable neutron-trunk service https://review.opendev.org/c/openstack/grenade/+/811117 | 17:28 |
| opendevreview | Ghanshyam proposed openstack/devstack stable/wallaby: Enable neutron trunk extension https://review.opendev.org/c/openstack/devstack/+/811457 | 17:30 |
| gmann | yeah ceph job is failing on stable/wallaby too https://zuul.opendev.org/t/openstack/build/1232607e55ff48c0b89d5e5b6fb4cfd4 | 17:40 |
| gmann | removing it | 17:40 |
| opendevreview | Ghanshyam proposed openstack/devstack stable/wallaby: Fix uwsgi config for trailing slashes https://review.opendev.org/c/openstack/devstack/+/811445 | 17:41 |
| frickler | I was wanting to wait for the results for xena, too, to see if other jobs might be broken there, too. likely at least the ceph job | 17:54 |
| frickler | for master only grenade was failing, so I didn't push that yet | 17:55 |
| clarkb | got it | 17:55 |
| opendevreview | Dr. Jens Harbott proposed openstack/devstack stable/xena: Fix uwsgi config for trailing slashes https://review.opendev.org/c/openstack/devstack/+/811389 | 18:03 |
| gmann | I am filling both bugs for tracking 1. uwsgi setup 2. ceph job broken so that it can be added back when fixed | 18:04 |
| opendevreview | Dr. Jens Harbott proposed openstack/devstack master: Fix uwsgi config for trailing slashes https://review.opendev.org/c/openstack/devstack/+/811303 | 18:07 |
| opendevreview | Archit Modi proposed openstack/whitebox-tempest-plugin master: Add Rbd direct download test https://review.opendev.org/c/openstack/whitebox-tempest-plugin/+/786979 | 18:31 |
| amodi | artom: ^^ im not sure how to create a separate periodic job https://review.opendev.org/c/openstack/whitebox-tempest-plugin/+/786979/36/.zuul.yaml | 18:32 |
| melwitt | clarkb: I'm not surprised to learn I got it all wrong 😝 sorry. thanks for explaining how it works. devstack-plugin-ceph used to be branchless and then branching was added later (starting in ussuri, I think). I'll work on finding if/how to untangle old stuff that was appropriate for branchless but is not good to have now | 18:34 |
| gmann | melwitt: it should pick up the job variant from stable/victoria which i fixed in https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/786309 | 18:50 |
| gmann | checking where things are going wring | 18:50 |
| gmann | wring | 18:50 |
| gmann | wrong | 18:50 |
| clarkb | gmann: I left detailed notes on the oen change showing why it happened | 18:51 |
| gmann | k | 18:51 |
| gmann | stable/xena is not setup properly as it is recently cut, let me remove it from master too for future release | 18:52 |
| melwitt | even better if gmann can fix it :) | 18:52 |
| gmann | I did not think of future stable branch that time we made this plugin branched | 18:53 |
| melwitt | gmann: here's a link to clarkb's notes https://review.opendev.org/c/openstack/devstack/+/811399/2#message-612206d57915d4370d0dc77dc7db92cd75865e26 | 18:54 |
| gmann | yeah, it is matching here https://zuul.opendev.org/t/openstack/build/0a01561f56414ffe8b2f659370c27666/log/zuul-info/inventory.yaml#241 | 18:54 |
| gmann | and yoctozepto mentioned it while I did cleanup few other stable branch but I thought of current stable branch only not future - https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/786307/1/.zuul.yaml | 18:56 |
| artom | amodi, eh, leave it out then, add a #TODO, and we can add it later | 19:01 |
| opendevreview | Ghanshyam proposed openstack/devstack-plugin-ceph master: Fix job definition as devstack-plugin-ceph is branched https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/811478 | 19:27 |
| opendevreview | Ghanshyam proposed openstack/devstack-plugin-ceph master: Fix job definition as devstack-plugin-ceph is branched https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/811478 | 19:29 |
| opendevreview | Ghanshyam proposed openstack/devstack-plugin-ceph master: Fix job definition as devstack-plugin-ceph is branched https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/811478 | 19:31 |
| opendevreview | Ghanshyam proposed openstack/devstack-plugin-ceph stable/xena: Fix job definition as devstack-plugin-ceph is branched https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/811348 | 19:33 |
| opendevreview | Ghanshyam proposed openstack/devstack stable/xena: Re-enable nova-ceph-multistore job https://review.opendev.org/c/openstack/devstack/+/811480 | 19:43 |
| gmann | melwitt: https://review.opendev.org/q/I7061f8d1491ff957452c9c777e40186a4e9c324e | 19:47 |
| gmann | testing it on devstack by re-enabling ceph job https://review.opendev.org/c/openstack/devstack/+/811480 | 19:48 |
| melwitt | thanks! will review | 19:48 |
| gmann | melwitt: also this one https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/810059 | 19:48 |
| melwitt | ack | 19:48 |
| *** amodi_ is now known as amodi | 19:50 | |
| opendevreview | Archit Modi proposed openstack/whitebox-tempest-plugin master: Add Rbd direct download test https://review.opendev.org/c/openstack/whitebox-tempest-plugin/+/786979 | 19:52 |
| gmann | now nova-grenade-multinode failure | 19:53 |
| gmann | clarkb: would you like to hit +A on this https://review.opendev.org/c/openstack/devstack/+/811303 | 19:54 |
| opendevreview | Ghanshyam proposed openstack/grenade master: zuul: Add the neutron devstack plugin and enable neutron-trunk service https://review.opendev.org/c/openstack/grenade/+/811117 | 20:01 |
| clarkb | gmann: done | 20:04 |
| *** elodilles is now known as elodilles_pto | 20:52 | |
| opendevreview | Ghanshyam proposed openstack/devstack stable/wallaby: Enable neutron trunk extension https://review.opendev.org/c/openstack/devstack/+/811457 | 21:02 |
| opendevreview | Ghanshyam proposed openstack/devstack stable/xena: Cap Volume, Swift, and Volume API extensions list for stable/xena https://review.opendev.org/c/openstack/devstack/+/811485 | 21:32 |
| gmann | lyarwood: gibi ^^ enabling neutron 'trunk' extension in stable/wallaby and stable/xena which should start capturing nova-grenade-multinode job failure and then we can test that with backporting the grenade fix | 21:38 |
| opendevreview | Merged openstack/devstack stable/victoria: Fix uwsgi config for trailing slashes https://review.opendev.org/c/openstack/devstack/+/811399 | 22:13 |
| opendevreview | Merged openstack/devstack stable/wallaby: Fix uwsgi config for trailing slashes https://review.opendev.org/c/openstack/devstack/+/811445 | 22:13 |
| opendevreview | Ghanshyam proposed openstack/devstack stable/wallaby: Re-enable nova-ceph-multistore job https://review.opendev.org/c/openstack/devstack/+/811490 | 22:21 |
| opendevreview | Merged openstack/devstack stable/xena: Fix uwsgi config for trailing slashes https://review.opendev.org/c/openstack/devstack/+/811389 | 22:28 |
| opendevreview | Merged openstack/devstack master: Fix uwsgi config for trailing slashes https://review.opendev.org/c/openstack/devstack/+/811303 | 22:28 |
| opendevreview | Ghanshyam proposed openstack/devstack stable/victoria: Re-enable nova-ceph-multistore job https://review.opendev.org/c/openstack/devstack/+/811489 | 22:34 |
| clarkb | gmann: I think https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/811478 shouldn't have zuul.d/stable-jobs in it | 22:36 |
| clarkb | gmann: the problem with those two chagnes is you still define the the jobs for stable/old in master and xena. Only one should do it. If master does it then every time you branch master you'll have cleanup to do | 22:38 |
| clarkb | My suggestion would be to put them on the oldest branch actually and also remove them from xena | 22:38 |
| clarkb | then when your oldest branch goes away the jobs definitions for the even older branches go away too | 22:38 |
| gmann | clarkb actually we have not cut the devstack-plugin-ceph branch for stabel/rocky or older and they are using jobs from master repo only. exlaining it in https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/811478/3/zuul.d/stable-jobs.yaml#1 | 22:38 |
| gmann | in this case I am just keeping for stable/rocky and older branch variant and for other stable where we have devstack-plugin-ceph branched, I have removed all branch variant in master | 22:39 |
| clarkb | but that isn't true because https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/811348/1/zuul.d/stable-jobs.yaml exists | 22:39 |
| clarkb | and when you branch master again you'll have the same problem all over again | 22:39 |
| clarkb | https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/811478/3/zuul.d/stable-jobs.yaml also has branch variants | 22:40 |
| gmann | is it? jobs there is only matching the stable/rocky and older | 22:40 |
| clarkb | yes but you match it in both. Its the same issue we fixed last week wtih devstack and the mkswap | 22:40 |
| clarkb | when you try to run a job on the old branch it is going to dobule up the pre-runs | 22:41 |
| clarkb | and the nfail because mkswap won't work | 22:41 |
| gmann | humm | 22:41 |
| clarkb | you should only ever use the branch: specification on a single branch | 22:41 |
| clarkb | my suggestion is whatever the oldest branch in that repo is so that they can go away when that branch goes away | 22:41 |
| gmann | this is special case where we have not branched stable/rocky or older | 22:41 |
| clarkb | right so on stable/stein define the jobs for stbale/rocky and older and only define them there | 22:42 |
| gmann | clarkb: yeah we can do that. so mkswap issue would not come there ? | 22:42 |
| clarkb | it shouldn't if you define the stable jobs for rocky and older on one branch only | 22:42 |
| gmann | ah right. | 22:42 |
| clarkb | and doing it on stein means that when stein is deleted we clean up the jobs for rocky and older | 22:43 |
| gmann | let me put those in stable/stein | 22:43 |
| gmann | I forgot to remove those from stable/xena backport but yes again it will be extra tasks to remove those for eveyr branch cut so it is better to move it from master | 22:44 |
| clarkb | exactly | 22:44 |
| opendevreview | Ghanshyam proposed openstack/devstack-plugin-ceph master: Fix job definition as devstack-plugin-ceph is branched https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/811478 | 22:47 |
| gmann | clarkb: ^^ | 22:47 |
| opendevreview | Ghanshyam proposed openstack/devstack-plugin-ceph master: Fix job definition as devstack-plugin-ceph is branched https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/811478 | 22:48 |
| clarkb | looks like you had stein and train specific jobs too? those should just be part of the branch | 22:48 |
| opendevreview | Ghanshyam proposed openstack/devstack-plugin-ceph master: Fix job definition as devstack-plugin-ceph is branched https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/811478 | 22:51 |
| gmann | clarkb: yeah we have there it just we forget to cleanup but let me double check | 22:51 |
| gmann | yeah they are there | 22:52 |
| opendevreview | Ghanshyam proposed openstack/devstack-plugin-ceph stable/stein: Moving the stable/rocky and odler stable jobs to stable/stein https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/811510 | 22:59 |
| opendevreview | Ghanshyam proposed openstack/devstack-plugin-ceph stable/stein: Moving the stable/rocky and odler stable jobs to stable/stein https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/811510 | 23:00 |
| opendevreview | Ghanshyam proposed openstack/grenade stable/xena: zuul: Add the neutron devstack plugin and enable neutron-trunk service https://review.opendev.org/c/openstack/grenade/+/811511 | 23:04 |
| opendevreview | Ghanshyam proposed openstack/grenade stable/wallaby: zuul: Add the neutron devstack plugin and enable neutron-trunk service https://review.opendev.org/c/openstack/grenade/+/811512 | 23:06 |
| opendevreview | Ghanshyam proposed openstack/devstack-plugin-ceph master: Fix job definition as devstack-plugin-ceph is branched https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/811478 | 23:09 |
| gmann | stable/xena is not setup yet, need to wait for this https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/811074/1 | 23:12 |
| gmann | gibi: lyarwood this is complete set og nova-grenade-multinode fix backport and testing patches - https://review.opendev.org/q/topic:%22bug%252F1945346%22+(status:open%20OR%20status:merged) | 23:27 |
| gmann | on testing patches, making sure this job is failing when trunk extension is enabled after that I will add depends-on: <grenade fix> on devstack patches and then recheck on testing patch | 23:28 |
| opendevreview | Ghanshyam proposed openstack/devstack-plugin-ceph stable/stein: Moving the stable/rocky and older stable jobs to stable/stein https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/811510 | 23:59 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!