| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2021-09-28 | |||
| 10:59:23 | opendevreview | Balazs Gibizer proposed openstack/nova master: Reproduce bug 1945310 https://review.opendev.org/c/openstack/nova/+/811394 | |
| 11:00:39 | opendevreview | Balazs Gibizer proposed openstack/nova master: Query ports with admin client to get resource_request https://review.opendev.org/c/openstack/nova/+/811396 | |
| 11:07:09 | lyarwood | gibi: I'm going to hold off posting to the thread, I think people are likely aware of the stable requirement here | |
| 11:07:34 | lyarwood | gibi: but I'm slightly confused by the way forward, it seems that people want to add the trailing / but that's breaking keystone at least on stable/xena | |
| 11:13:32 | gibi | lyarwood: ohh, I haven't checked the stable/xena result yet | |
| 11:17:22 | gibi | lyarwood: do you think a non trailing / version would pass? | |
| 11:17:40 | lyarwood | gibi: it passed on master for the non-grenade jobs | |
| 11:17:46 | gibi | hm, true | |
| 11:18:08 | gibi | so then we have to go back and reason about the non trailing / case | |
| 11:24:56 | lyarwood | yup I think so | |
| 12:13:17 | opendevreview | Balazs Gibizer proposed openstack/nova master: Query ports with admin client to get resource_request https://review.opendev.org/c/openstack/nova/+/811396 | |
| 12:19:55 | opendevreview | Balazs Gibizer proposed openstack/nova stable/xena: Reproduce bug 1945310 https://review.opendev.org/c/openstack/nova/+/811405 | |
| 12:21:11 | opendevreview | Balazs Gibizer proposed openstack/nova stable/xena: Query ports with admin client to get resource_request https://review.opendev.org/c/openstack/nova/+/811407 | |
| 12:28:58 | gibi | bauzas: now I can confirm that https://bugs.launchpad.net/nova/+bug/1945310 is not fully a xena regression, at least parts of it is reproducible in wallaby too | |
| 12:29:17 | bauzas | gibi: ack, then remove the xena regression tag | |
| 12:29:49 | bauzas | gibi: can you backport the reproduction change to stable/wallaby so the CI would test it ? | |
| 12:30:04 | gibi | I'm doing it right now | |
| 12:30:06 | bauzas | cool | |
| 12:30:13 | bauzas | so we don't need a RC3 for it | |
| 12:30:15 | gibi | this is how I confirmed that it is a probelm in wallaby too | |
| 12:30:31 | bauzas | but we will merge it for a 25.0.1 release | |
| 12:30:37 | bauzas | huh | |
| 12:30:40 | bauzas | 24.0.1 | |
| 12:30:45 | gibi | :) | |
| 12:31:04 | bauzas | gibi: thanks for working on it, very appreciated | |
| 12:31:34 | gibi | I broke it :D | |
| 12:31:44 | gibi | so hell yeah I have to work on it | |
| 12:47:50 | jamesdenton_alt | Any ideas as to why Nova might return duplicates for hypervisors and services (same UUIDs)? https://paste.opendev.org/show/809646/ | |
| 13:16:17 | opendevreview | Ivan Kolodyazhny proposed openstack/nova master: Fix NVMe volume attach https://review.opendev.org/c/openstack/nova/+/811413 | |
| 13:20:20 | opendevreview | Balazs Gibizer proposed openstack/nova stable/wallaby: Reproduce bug 1945310 https://review.opendev.org/c/openstack/nova/+/811414 | |
| 13:21:35 | gibi | bauzas: so on stable/wallaby we have 2/3 of the bug exists ^^ | |
| 13:21:38 | opendevreview | Balazs Gibizer proposed openstack/nova stable/wallaby: Query ports with admin client to get resource_request https://review.opendev.org/c/openstack/nova/+/811416 | |
| 13:22:17 | bauzas | gibi: over the phone for a meeting but ack | |
| 13:22:56 | gibi | me too | |
| 13:24:54 | opendevreview | Ivan Kolodyazhny proposed openstack/nova master: Fix NVMe volume attach https://review.opendev.org/c/openstack/nova/+/811413 | |
| 13:31:35 | opendevreview | Balazs Gibizer proposed openstack/nova stable/wallaby: Query ports with admin client to get resource_request https://review.opendev.org/c/openstack/nova/+/811416 | |
| 13:58:04 | zigo | Looks like nova-api on Xena (over Debian Bullseye) doesn't like me... | |
| 13:58:16 | zigo | Failed to delete allocations for consumer 26ee7dd9-75c2-41f5-86f4-c714ed36aab7. Error: {"errors": [{"status": 500, "title": "Internal Server Error", "detail": "The server has either erred or is incapable of performing the requested operation.\n\n unsupported callable ", "request_id": "req-34a9215b-9387-43b0-ae24-2d85c7d2b8f5"}]} (HTTP 409) (Request-ID: req-57d65ddf-f845-4fa4-9f13-4382ec19dc72) | |
| 13:58:16 | zigo | # openstack server delete 26ee7dd9-75c2-41f5-86f4-c714ed36aab7 | |
| 13:58:55 | zigo | When doing so, nova-api just spams its log with some policy deprecation stuff... | |
| 13:59:07 | zigo | How am I supposed to generate the default policy file?!? | |
| 13:59:42 | zigo | This looks like completely broken AGAIN... :( | |
| 14:01:10 | zigo | Also, has the /healthcheck changed behavior?!? | |
| 14:02:38 | zigo | Now it wants me to be authenticated ?!? | |
| 14:02:44 | zigo | Seriously? | |
| 14:03:04 | zigo | How am I supposed to do that with the standard haproxy thingy? | |
| 14:03:25 | bauzas | afaik we haven't changed the paste.ini | |
| 14:03:40 | bauzas | the whole policy thing is different | |
| 14:04:02 | bauzas | but i need to go away or I'll be pedestrian | |
| 14:05:38 | zigo | When I do a: curl -k https:// |
|
| 14:05:49 | zigo | I'm not even sure that can be done with haproxy. | |
| 14:05:49 | opendevreview | Ivan Kolodyazhny proposed openstack/nova master: Fix NVMe volume attach https://review.opendev.org/c/openstack/nova/+/811413 | |
| 14:08:12 | zigo | It looks like its probably my bad, checking ... | |
| 14:08:24 | gibi | lyarwood: does this ring a bell to you https://review.opendev.org/c/openstack/devstack/+/811399/1#message-941958791253e3555a648888281a5de6cbf3b1c4 ? | |
| 14:12:39 | lyarwood | gibi: nope never seen that before | |
| 14:12:58 | lyarwood | http://review.opendev.org/ dead for everyone? | |
| 14:14:29 | gibi | works for me | |
| 14:14:33 | kashyap | Yeah, same here | |
| 14:14:36 | dansmith | same | |
| 14:14:38 | kashyap | (As in, works) | |
| 14:15:43 | dansmith | bauzas: is there a tentative ptg schedule yet? | |
| 14:17:42 | gibi | dansmith: we have two cross project discussions scheduled to a certain slot but top of that we haven't assigned topics to slots yet | |
| 14:17:54 | gibi | https://etherpad.opendev.org/p/nova-yoga-ptg | |
| 14:18:05 | gibi | lyarwood: ack, then I recheck :) | |
| 14:18:47 | dansmith | gibi: okay, it's always that arms race where everyone is trying to avoid others' sessions, but nobody has scheduled yet :) | |
| 14:24:16 | gibi | I think it is also part of not having the PTG planning in focus until after Xena is out of the tood | |
| 14:24:19 | gibi | door | |
| 14:24:49 | zigo | https://paste.opendev.org/show/809649/ | |
| 14:24:49 | zigo | When trying to delete a server that went into error, I get this in placement.log: | |
| 14:24:56 | zigo | Any idea why? | |
| 14:25:08 | dansmith | I dunno, I think the "I don't want to schedule X until I know of person Y is available, but person Y's schedule depends on other project Z" problem would exist either way | |
| 14:26:26 | gibi | dansmith: true | |
| 14:26:30 | gibi | zigo: interesting | |
| 14:26:49 | zigo | Maybe I should try downgrading to oslo.policy 3.7.0 to see what's going on? | |
| 14:26:56 | dansmith | that error can come from the json encoder/decoder, maybe there's something wrong with the policy file? | |
| 14:27:26 | zigo | dansmith: What I'm surprised of, is that I'm using a yaml policy file in /etc/placement/policy.d/00_default_policy.yaml | |
| 14:27:43 | dansmith | zigo: ah, good, that was going to be my next question | |
| 14:27:54 | zigo | And there, everything was supposed to be commented out, but it's not: every DEPRECATED stuff is active ... | |
| 14:28:00 | zigo | Same for Nova. | |
| 14:28:07 | zigo | (I had to comment out everything...) | |
| 14:28:51 | zigo | dansmith: So you can rule out a policy.json thingy, as I don't have one ! :) | |
| 14:30:39 | zigo | Same result with oslo.policy 3.7.0 | |
| 14:33:23 | zigo | Ok, got it to work ... | |
| 14:33:42 | zigo | Again, same stuff: oslopolicy-sample-generator is generating a default policy yaml file that DOES NOT WORK ! | |
| 14:33:52 | zigo | When commenting everything out, it works... | |
| 14:34:20 | gmann | zigo: I do not think oslopolicy-sample-generator add the deprecated rules | |
| 14:34:41 | gmann | zigo: best way is not to have policy file if no rule overridden is needed | |
| 14:34:41 | zigo | gmann: That's what it did for me ... | |
| 14:35:26 | zigo | gmann: Well, in my package, I'm writing a /etc/placement/policy.d/00_default_policy.yaml with the expectation that this fille will only contain commented out directives ... | |
| 14:35:33 | zigo | It worked this way for the last 2 or 3 releases. | |
| 14:35:36 | zigo | Now that's broken. | |
| 14:36:25 | gmann | zigo: commented out you mean except deprecated rules right? | |
| 14:36:43 | zigo | gmann: Why would there be some special cases for the deprecated rules? | |
| 14:37:32 | gmann | zigo: no, I mean everything was commented out ? because earlier you mentioned deprecated rules were not commented? | |
| 14:38:13 | zigo | gmann: That's what I mean yes, in my packages, now, I have everything commented out but all what's deprecated ... :/ | |
| 14:39:18 | gmann | I feel we should extend these tool oslopolicy-sample-generator and oslopolicy-policy-generator to add deprecated rule too with flag so that it does not break deployment not moved to new policies yet | |
| 14:40:10 | zigo | Anyway, GOT MY FIRST Xena VM UP AND RUNNING AND PINGING !!! \o/ | |
| 14:40:28 | gibi | \o/ | |
| 14:40:28 | gmann | zigo: can you paste your old (failing) policy file? | |
| 14:40:32 | gmann | zigo: perfect | |