Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2024-09-18
07:47:53 frickler thx, I commented on this yesterday already in the qa channel, I was under the impression that sean-k-mooney had tested my change, sorry for that. is this option also really named differently between this and what create and rebuild use? might be just an sdk thing, but certainly not helpful
07:48:25 frickler I'm also not really here this week, so try to ping other sdk, release and reqs people as needed
08:12:50 bauzas sorry folks, I was off the laptop last hour
08:12:52 bauzas now I'm back
08:14:39 bauzas so, I'd prefer to rollback the u-c by OSC6 now
08:16:16 bauzas gibi: elodilles ^
08:16:19 bauzas stephenfin: too ^
08:40:50 danfai_ Quick question: In nova I saw that the name is sanitized to be used as a hostname. The hostname should then be maximum 63 characters long. In case you use a name that is longer, the last characters are truncated. If you now use the server create call to create multiple VMs, this length will be longer, since the -{index} is appended. Is this supposed to be like this?
09:30:09 gibi bauzas: do you suggest to reopen https://review.opendev.org/c/openstack/requirements/+/929552 ? I have no power to un-abandon it.
09:37:05 sean-k-mooney i can do that now
09:37:43 sean-k-mooney done
09:37:50 sean-k-mooney although it needs a rebase
09:37:56 sean-k-mooney will i do that now
09:38:05 sean-k-mooney and add 7.1.2?
09:39:03 sean-k-mooney sorry 7.1.1
09:40:20 gibi yeah 7.1.1 is still broken
09:42:29 sean-k-mooney ok updated https://review.opendev.org/c/openstack/requirements/+/929552
09:43:59 gibi sean-k-mooney: thanks
09:44:23 gibi I pinged the release channel as well
09:45:16 sean-k-mooney looking back at my orgial comment i linked to the code https://review.opendev.org/c/openstack/python-openstackclient/+/929236/comment/d39df785_1f261b9e/ to check the parmater was correct
09:45:33 sean-k-mooney and i misread admin_password as admin_pass
09:51:05 bauzas sorry, I was again off (I have a tech folk in my house for my optic fiber issue)
09:51:51 bauzas sean-k-mooney: stephenfin: I'm really sorry about breaking the OSC7 version, but we can't wait for a new release for like 2 days
09:52:20 bauzas and given https://bugs.launchpad.net/nova/+bug/2081023 I'd prefer to have time for testing it
10:16:23 opendevreview sean mooney proposed openstack/nova master: [DNM] testing unreleased osc https://review.opendev.org/c/openstack/nova/+/929757
10:17:45 sean-k-mooney that shoudl pull osc master
10:18:07 sean-k-mooney im going to push another dnm to instead use the pinned version of osc
10:19:34 opendevreview sean mooney proposed openstack/nova master: [DNM] testing pinning osc to 6.6.0 https://review.opendev.org/c/openstack/nova/+/929758
11:30:03 gibi proposed osc 7.1.2 so that if the decision is to roll forward then we can have a place to roll to https://review.opendev.org/c/openstack/releases/+/929765
11:41:29 bauzas sean-k-mooney: have you seen gibi's https://bugs.launchpad.net/nova/+bug/2081023 when testing evacuate with my patch ?
11:42:11 sean-k-mooney i saw it but didnt look at it in detail
11:42:13 bauzas gibi: I'm not super happy with the move forward now, when trying to understand how OSC was calling the SDK, I wasn't sure my patch was the right thing to do
11:42:39 sean-k-mooney the issue is the sdk renames all the api field internally
11:42:48 sean-k-mooney and that mapping is really easy to get wrong
11:43:17 gibi bauzas: I'm not advocationg to roll forward any more. I just preparing for the chance that the decision is to roll forward
11:43:30 gibi also 7.1.2 needed anyhow to get a working osc out to the users
11:44:01 gibi bauzas: also https://bugs.launchpad.net/nova/+bug/2081023 is not osc but nova
11:44:08 bauzas I think we need time to test OSC7
11:44:10 bauzas that's the point
11:44:11 gibi and only cosmetic as far as I see
11:44:12 sean-k-mooney i have two sets of zuul jobs usin gnova gate with master osc and rolling back to 6.6.0
11:44:16 bauzas while RC1 is now super urgent
11:44:42 sean-k-mooney and we have the patches to requireemtn repo to go either direction
11:44:44 gibi bauzas: 7.1.1 is out there as we released it, so having a fixed 7.1.2 out there helps
11:44:47 gibi the users
11:44:49 sean-k-mooney so we can use that to inform the decsion
11:45:52 bauzas gibi: well, I wouldn't say that #2081023 is cosmetic
11:46:29 bauzas as an admin, you either need to set older microversion or upgrade your computes in order to not get that failure, right?
11:48:27 sean-k-mooney are we sure this is a bug
11:48:29 gibi bauzas: it happens because all of the nodes are forced down in the deployment
11:48:41 gibi and that results in min service version to be 0
11:48:52 sean-k-mooney this is directly caused by osc using the latest microversion is it not?
11:48:53 bauzas right
11:48:55 gibi but if all nodes are down then evacuation would fail anyhow
11:49:06 gibi sean-k-mooney: it is independent from osc changes
11:49:09 bauzas correct but we get that super weird exception
11:49:18 gibi it is in nova since 2.95
11:49:32 gibi jepp the exception test needs the cosmetic surgery
11:49:34 gibi that is all
11:49:40 gibi *text
11:50:07 gibi I just noticed it during testing the 7.1.2 fix in a single node devstack
11:50:44 gibi so I had to user <2.95 in the test hence mentioning why I needed that
11:50:47 gibi *use
11:51:01 sean-k-mooney ok so this is not a blocker to rollforward to 7.1.2 then as its not related
11:51:17 gibi it is not a blocker at all in RC1 and it is not related to osc at all
11:51:29 sean-k-mooney its just a bad error message
11:51:33 gibi yes
11:51:41 gibi as far as I see it is only a bad error message
11:51:59 sean-k-mooney bauzas: for context people kept asking osc to make nova ues the latest microverion
11:52:00 gibi and a bit hard to understand behavior around forced down nodes and service version
11:52:20 sean-k-mooney this is what happens when you do that with an older cloud it will use a microversoin that may not be supported
11:52:25 bauzas sean-k-mooney: yup, and I'm on the same path
11:52:33 bauzas I prefer our users to call the latest
11:52:47 sean-k-mooney well i really dont think that was the right path
11:52:57 bauzas only CLI and scripts should pin
11:53:02 sean-k-mooney osc shoudl use the oldest microverison supproted by the current command
11:53:06 bauzas for stability reasons
11:53:17 gibi (I saw another force down issue, with a force down / up I can force a not running compute to be reported as up but I had not time to file it. Probably not an RC regression either)
11:53:29 bauzas (when I say CLI, I mean programmatical usage of the CLI)
11:53:53 sean-k-mooney whihc is officaly not a supproted usecase of osc now
11:54:08 sean-k-mooney osc does not commit to providign a programable srcipting interface
11:54:20 sean-k-mooney whcih is explcitly why the puppet modules shoudl not use it oging forward
11:54:23 bauzas then, I'm definitely in favor of calling latest
11:54:48 bauzas as a user, I shouldn't opt into new features
11:55:32 bauzas man, the RC1 issue prevents me to put full attention on something
14:09:24 opendevreview Rajesh Tailor proposed openstack/nova-specs master: Show finish_time field in instance action show https://review.opendev.org/c/openstack/nova-specs/+/929780
14:25:43 opendevreview Rajesh Tailor proposed openstack/nova-specs master: Show finish_time field in instance action show https://review.opendev.org/c/openstack/nova-specs/+/929780
17:38:11 auniyal dansmith, Format inspector failed, aborting: Signature KDMV not found: b'QFI\xfb' it should be fixed by https://review.opendev.org/c/openstack/glance/+/927291 right
17:38:47 dansmith no?
17:39:03 auniyal I am getting this from nova while creating a guest from new ubuntu image, from here https://docs.openstack.org/image-guide/obtain-images.html#ubuntu
17:39:03 auniyal and I
17:39:24 dansmith that error means something is trying to validate a vmdk but likely found a qcow file or something
17:40:21 auniyal I downloaded .img and created glance image in qcow2 format with cmd: openstack image create --disk-format qcow2 --file file.img ubuntu
17:41:09 auniyal should I need to do something first, before image create ?
17:41:25 dansmith no
17:41:42 dansmith is the file actually a qcow?
17:42:29 auniyal I tried this, it says qcoe
17:42:31 auniyal file noble-server-cloudimg-amd64.img
17:42:31 auniyal noble-server-cloudimg-amd64.img: QEMU QCOW2 Image (v3), 3758096384 bytes
17:43:08 auniyal s/says qcow/qcoe

Earlier   Later