| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2019-09-25 | |||
| 13:32:05 | stephenfin | (I already have a glossary at the top of the document explaining the terms) | |
| 13:32:30 | bauzas | meaning that I'm fine with +2ing the change and just saying that we could provide better explanations in some follow-up | |
| 13:32:54 | bauzas | stephenfin: eh ok | |
| 13:33:29 | bauzas | FWIW, my personal opinion is that we should just accept any documentation providing more explanations and just asking for follow-ups if some nits are there | |
| 13:33:53 | bauzas | unless a doc is wrong, honestly no need for -1ing it | |
| 13:34:11 | bauzas | because it's just better to have a documentation than none | |
| 13:35:59 | mriedem | kashyap: fyi if you or any other libguestfs minded people are about https://bugs.launchpad.net/nova/+bug/1845284 | |
| 13:35:59 | openstack | Launchpad bug 1845284 in OpenStack Compute (nova) "Config CONF.libvirt.inject_partition=-1, build instance with metadata will remain some images in /var/tmp/.guestfs-0/appliance.d" [Low,New] | |
| 13:36:08 | kashyap | mriedem: Hiya | |
| 13:36:10 | openstackgerrit | Stephen Finucane proposed openstack/nova master: docs: Document how to revert, confirm a cold migration https://review.opendev.org/664396 | |
| 13:36:21 | gibi | bauzas: hi! is there a way to fake some VGPUs in devstack without having real hardware? | |
| 13:36:28 | openstackgerrit | Stephen Finucane proposed openstack/nova master: docs: Update resize doc https://review.opendev.org/665412 | |
| 13:36:34 | bauzas | gibi: hah, good question | |
| 13:37:06 | stephenfin | bauzas: In that case, I have two lovely docs here for you ;) https://review.opendev.org/#/c/670125/ https://review.opendev.org/#/c/667165/ | |
| 13:37:13 | stephenfin | assuming you know anything about quotas | |
| 13:37:26 | stephenfin | bauzas: and if not, https://review.opendev.org/#/c/667133/ is definitely your wheelhouse | |
| 13:38:01 | bauzas | stephenfin: I still have one from you I remember to review https://review.opendev.org/#/c/667133/ | |
| 13:38:08 | bauzas | hah, this | |
| 13:38:23 | kashyap | mriedem: Right off the bat, the "No operating system ..." error from libguestfs means what it says on the tin - libguestfs couldn't find an OS in the disk - a bogus disk, corrupted, etc | |
| 13:38:24 | stephenfin | That's the one :D | |
| 13:38:33 | mriedem | lyarwood: melwitt: mnaser: stephenfin: weee ironic serial consoles are broken https://bugs.launchpad.net/nova/+bug/1845243 | |
| 13:38:33 | openstack | Launchpad bug 1845243 in OpenStack Compute (nova) "Nested 'path' query param in console URL breaks serialproxy" [Undecided,New] | |
| 13:38:36 | kashyap | mriedem: I'll add words to the bug | |
| 13:38:42 | mriedem | kashyap: thanks | |
| 13:39:20 | bauzas | gibi: I was thinking of using the fake libvirt driver (https://libvirt.org/drvtest.html) for telling that we have mdevs, but looks like we have problems with it | |
| 13:39:37 | mriedem | stephenfin: i reviewed https://review.opendev.org/#/c/670125/ once, just forgot about it, will look again | |
| 13:39:46 | stephenfin | mriedem: That's unfortunate. Should be an easy fix though. Lemme try | |
| 13:39:48 | gibi | bauzas: thanks | |
| 13:39:57 | bauzas | gibi: for example, I could pass my own XML to the test driver | |
| 13:39:59 | kashyap | mriedem: Also, what they're asking is completely off: if I parse the reporter, they want the "images in the /var/tmp/ files to be "left forever" | |
| 13:40:22 | kashyap | ... except that's by design; the hint is in the file path, they _are_ temporary images that _will_ go away. | |
| 13:41:26 | mriedem | kashyap: i think they are saying the images are left and should be removed | |
| 13:41:42 | mriedem | expected: "In the end of booting instance, can remove those images in /var/tmp/.guestfs-0/appliance.d" | |
| 13:41:48 | mriedem | but yeah, if it's /tmp they should be gone on restart | |
| 13:41:58 | kashyap | mriedem: I parsed it upside down; by design of my brain; ignore me | |
| 13:42:34 | kashyap | But guestfs / libvirt / QEMU version is needed to reproduce. And you asked it there | |
| 13:49:05 | kashyap | mriedem: So, checking w/ a libguestfs maintainer (Rich Jones) -- he says files in that /var/tmp are a cache and libguestfs _won't_ remove them | |
| 13:49:13 | kashyap | It relies on external component to do it | |
| 13:49:27 | kashyap | And Nova _shouldn't_ delete them either - that feels wrong. | |
| 13:49:53 | mriedem | kashyap: ok, so basically it's glance but at a lower level :) | |
| 13:50:14 | mriedem | i seem to remember issues in ibm's internal CI where booting the first guest would take longer b/c libguestfs had to build the cache for the image | |
| 13:50:17 | mriedem | something like that | |
| 13:50:18 | kashyap | Yeap | |
| 13:50:37 | kashyap | mriedem: Also there's a systemd tool (there are a gazillion of them) which removes temp files, while making a sandwich for you | |
| 13:51:08 | kashyap | Ha! Found the sucker: | |
| 13:51:14 | kashyap | "`systemd-tmpfiles` creates, deletes, and cleans up volatile and temporary files and directories, based on the configuration file format and location specified in tmpfiles.d(5)" | |
| 13:51:32 | kashyap | So they should something like that | |
| 13:52:23 | mriedem | kashyap: ok i guess just drop those details in the bug report and then mark it invalid | |
| 13:52:26 | mriedem | thanks | |
| 13:52:32 | kashyap | Also, unrelated funninness (is it a word?) I have to admit: "the feast of maximum strikage" :D | |
| 13:54:51 | kashyap | mriedem: Done; words added, turned the state to invalid. | |
| 13:55:43 | mriedem | TheJulia: do you know if it's possible to test ironic serial consoles in tempest? https://bugs.launchpad.net/nova/+bug/1845243 | |
| 13:55:43 | openstack | Launchpad bug 1845243 in OpenStack Compute (nova) stein "Nested 'path' query param in console URL breaks serialproxy" [High,Confirmed] | |
| 13:55:57 | mriedem | tempest has a test for novnc consoles but i'm not sure how possible that is with ironic serial consoles | |
| 14:01:56 | mriedem | gibi: i feel like i should get some retroactive compensation from ericsson for getting the stable branch extended maintenance policy change enacted :) | |
| 14:02:04 | mriedem | making all of your pike dreams come true | |
| 14:02:37 | mriedem | i bet there is still someone at the ops meeting that wants to talk about LTS though... | |
| 14:02:49 | gibi | mriedem: Thank you. (I can give you the email of my boss. but I guess that will not help much. ) | |
| 14:03:01 | mriedem | ha | |
| 14:04:33 | gibi | and it would be so nice to have more people from ericsson contribution to the ext maintenance effort | |
| 14:05:06 | gibi | and I continously ask for more people without success | |
| 14:13:50 | efried | stephenfin: You had a -1 on a previous PS of https://review.opendev.org/#/c/683783/ -- I think your feedback was addressed (repath was reverted) but would you please have another quick look and make sure there's nothing blocking remaining? | |
| 14:14:55 | bauzas | gibi: mriedem: honestly, not a lot of people work for 100% upstream you know | |
| 14:15:08 | bauzas | so we have a lot of other stuff | |
| 14:15:09 | efried | sean-k-mooney: I'll proxy your approval of https://review.opendev.org/#/c/683437/ if you could have another look please ("everything CPU pinning" doc) | |
| 14:15:54 | bauzas | efried: like I said previously, unless a doc is having some wrong explanation, I should we should just accept it, and then providing some FUP | |
| 14:16:00 | bauzas | (in case of comments) | |
| 14:16:01 | gibi | bauzas: I would be happy to get an extra two half-people for stable/pike from my employee but... | |
| 14:16:56 | efried | bauzas: I agree; I'm trying to delegate to someone else to confirm that there's nothing actually wrong in there. | |
| 14:18:03 | bauzas | stephenfin: FWIW, finally looking at https://review.opendev.org/#/c/667133 :) | |
| 14:21:54 | mriedem | bauzas: i thought most people were 100% upstream... | |
| 14:22:15 | bauzas | yeah, me too honestly | |
| 14:22:32 | mriedem | sorry, i'm being facetious | |
| 14:22:40 | bauzas | and then I discovered the other way once I went a core | |
| 14:22:52 | bauzas | because 'heh, can you do this please' | |
| 14:22:59 | bauzas | or 'look at this, is this right?' | |
| 14:36:06 | mriedem | gibi: looks like the show version tempest test failures in https://review.opendev.org/#/c/684658/ are real (on pike) | |
| 14:36:24 | mriedem | tempest doesn't test pike since it's in extended maintenance mode so it's likely something changed in tempest | |
| 14:36:47 | mriedem | probably dropped support for cinder v1 or v2 or something | |
| 14:39:44 | mriedem | this https://review.opendev.org/#/c/680790/ | |
| 14:39:51 | mriedem | gmann: stable/pike is also broken since ^ | |
| 14:40:57 | mriedem | so i guess we're going to have to run with a tagged tempest on stable/pike as well | |
| 14:41:17 | openstackgerrit | Stephen Finucane proposed openstack/python-novaclient master: Stop silently ignoring invalid 'nova boot --hint' options https://review.opendev.org/684762 | |
| 14:41:31 | stephenfin | that novaclient unit tests are dumb and I hate them | |
| 14:42:33 | gibi | mriedem: thanks | |
| 14:42:41 | gibi | mriedem: I will have to get back to that | |
| 14:43:24 | mriedem | my guess is tempest is failing b/c the cinder v1 api was removed in queens https://review.opendev.org/#/c/499342/ | |
| 14:46:35 | gibi | I try to get elod look into this | |
| 14:51:24 | mriedem | gibi: just posted to the ML | |
| 14:54:02 | gibi | mriedem: ack, thanks | |
| 14:58:09 | mriedem | https://review.opendev.org/#/c/684769/ | |
| 15:03:16 | openstackgerrit | Martin Midolesov proposed openstack/nova master: Implementing graceful shutdown. https://review.opendev.org/666245 | |
| 15:06:14 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tests: Ignore warning from sqlalchemy-migrate https://review.opendev.org/684772 | |
| 15:06:15 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tests: Nuke OS_STDOUT_CAPTURE, OS_STDERR_CAPTURE https://review.opendev.org/684773 | |
| 15:06:15 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tox: Use common 'command' definition for unit tests https://review.opendev.org/684774 | |
| 15:06:16 | openstackgerrit | Stephen Finucane proposed openstack/nova master: tox: Stop overriding the 'install_command' https://review.opendev.org/684775 | |
| 15:06:30 | openstackgerrit | Balazs Gibizer proposed openstack/nova stable/pike: Delete instance_id_mappings record in instance_destroy https://review.opendev.org/684658 | |
| 15:06:32 | stephenfin | mriedem: You'd asked for the first of those last week ^ | |
| 15:07:19 | mnaser | mriedem: ouch, i think it should be easy to fix .. i hope :X | |
| 15:10:17 | mriedem | stephenfin: so about that, | |