Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2024-06-11
16:24:38 fwiesel How do I continue from there on?
16:25:10 fwiesel Register it right away and put that in the url?
16:25:32 fwiesel Or ask here for feedback? (as I have done then implicitly)
16:26:27 bauzas well, usually, we discuss specless blueprints in our meetings
16:26:40 bauzas if you want to have it accepted, it's the right time
16:26:47 bauzas or do you want to discuss the design ?
16:27:16 bauzas oh, that's a spec template, nevermind
16:27:44 bauzas I think you can now create the blueprint, link it in the spec and upload that spec
16:28:05 fwiesel in nova-specs then?
16:28:22 bauzas you need to register a blueprint with https://blueprints.launchpad.net/nova/+spec/lazy-metadata-loading
16:28:45 bauzas then you need to upload the file you created in the nova-specs/2024.2/approved repor
16:28:53 bauzas s/repor/directory
16:29:00 bauzas and then push that commit to gerrit
16:30:01 fwiesel Okay, thanks. Will do
16:32:08 bauzas fwiesel: pro-tip, in the commit msg (in the nova-specs repo), add a specific "blueprint lazy-metadata-loading" tag in the text
16:32:19 bauzas something like "Proposes blueprint lazy-metadata-loading"
16:32:30 bauzas it will automatically link the spec in the blueprint whiteboard
16:33:18 bauzas ideally, create local branch from that name, so gerrit will have this topic
16:33:53 fwiesel Great, I'll do that.
16:36:44 bauzas okay, anything else ?
16:40:45 bauzas ok, I think we can close the meeting
16:40:48 bauzas thanks guys
16:40:52 opendevmeet Log: https://meetings.opendev.org/meetings/nova/2024/nova.2024-06-11-16.00.log.html
16:40:52 opendevmeet Minutes (text): https://meetings.opendev.org/meetings/nova/2024/nova.2024-06-11-16.00.txt
16:40:52 opendevmeet Minutes: https://meetings.opendev.org/meetings/nova/2024/nova.2024-06-11-16.00.html
16:40:52 opendevmeet Meeting ended Tue Jun 11 16:40:52 2024 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)
16:40:52 bauzas #endmeeting
16:41:08 elodilles thanks o/
16:42:05 fwiesel \o
#openstack-nova - 2024-06-12
05:31:47 opendevreview Takashi Kajinami proposed openstack/nova master: WIP: Migrate MEM_ENCRYPTION_CONTEXT from root provider https://review.opendev.org/c/openstack/nova/+/921814
06:05:08 opendevreview Takashi Kajinami proposed openstack/nova master: WIP: Migrate MEM_ENCRYPTION_CONTEXT from root provider https://review.opendev.org/c/openstack/nova/+/921814
07:03:39 opendevreview Takashi Kajinami proposed openstack/nova master: WIP: Migrate MEM_ENCRYPTION_CONTEXT from root provider https://review.opendev.org/c/openstack/nova/+/921814
07:40:55 opendevreview Sahid Orentino Ferdjaoui proposed openstack/nova master: scheduler: fix _get_sharing_providers to support unlimited aggr https://review.opendev.org/c/openstack/nova/+/921665
12:49:36 opendevreview Sahid Orentino Ferdjaoui proposed openstack/nova master: scheduler: fix _get_sharing_providers to support unlimited aggr https://review.opendev.org/c/openstack/nova/+/921665
14:01:55 ozzzo_work We built a new venv for wallaby which uses openstack CLI version 6.6.0 (old venv used 5.5.0), and now I get an error when I try to use "openstack server list --limit -1" to get all hosts: https://paste.openstack.org/show/bYNdWz8brEgB2ovSGiUF/
14:02:11 ozzzo_work how can I get a list of all VMs in 6.6.0?
15:18:48 opendevreview Sahid Orentino Ferdjaoui proposed openstack/nova master: scheduler: fix _get_sharing_providers to support unlimited aggr https://review.opendev.org/c/openstack/nova/+/921665
15:19:43 opendevreview Sahid Orentino Ferdjaoui proposed openstack/nova master: scheduler: fix _get_sharing_providers to support unlimited aggr https://review.opendev.org/c/openstack/nova/+/921665
15:27:07 melwitt I have found that it's not possible to install the ceph reef release with devstack-plugin-ceph (due to a recent-ish change) and there's changes proposed that AFAICT would remove the ability to install ceph on ubuntu entirely https://review.opendev.org/q/topic:%22cleanup-ceph-installer%22
15:28:56 frickler melwitt: that change should only drop the old pkg based install and use cephadm instead? at least that's how I understood gouthamr's mail
15:29:46 gouthamr frickler: +1 ; melwitt: i use reef most of the time these days.. what's failing?
15:30:15 melwitt frickler: hm, so that may mean it's not doing so intentionally but AFAICT from the code, it would have that result https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/918952/6/devstack/lib/cephadm#141
15:31:20 frickler melwitt: according to the comment the binary should be the same for all distros
15:32:19 melwitt I only see it pulling rpm packages in that function? I dunno if I'm just totally misunderstanding this
15:33:10 frickler it is pulling the cephadm binary afaict, which then is executed to do everything else
15:35:29 gouthamr melwitt: yes that's confusing, i thought my comment above would help with the confusion
15:35:59 gouthamr melwitt: https://zuul.opendev.org/t/openstack/build/a9d3d319d36f403f9e33febec5064a48/log/controller/logs/devstacklog.txt#9960 -- here's a recent integrated gate/CI job deploying reef
15:36:29 melwitt hm, ok. so the curl -f -O https://download.ceph.com/rpm-${ceph_version}/${os_release}/noarch/cephadm should be a binary that also runs on ubuntu I guess?
15:36:33 gouthamr yes
15:36:55 gouthamr melwitt: its not an RPM even if its put, very confusingly, in a rpm repo
15:37:35 gouthamr (the integrated job above runs on ubuntu 22.04/jammy)
15:37:38 melwitt I ran with CEPHADM_DEPLOY=true yesterday and it installed quincy even though I used CEPH_RELEASE=reef so I guess I should have known that. so I guess the only issue currently is that it can't install reef for some reason
15:38:11 melwitt i.e. at the end of the run 'ceph version' showed me it installed quincy
15:38:16 gouthamr melwitt: that's something we can clean up.. i think the CONTAINER_IMAGE variable is pointing to quincy
15:38:49 gouthamr or not: https://github.com/openstack/devstack-plugin-ceph/blob/13f94aaaf2b4a59581b1be5979600ca18a8df5c3/devstack/lib/cephadm#L32
15:40:33 melwitt yeah, I saw that too. but it ended up installing quincy and I don't remember how. at the moment I'm running devstack-plugin-ceph off the unmaintained/zed branch with a couple of tweaks to make it accept "jammy" and it's successfully installed reef
15:41:39 melwitt actually let me pull up logs on our nova-ceph-multistore job to see if it shows there
15:45:08 melwitt it starts out pulling the reef container image https://zuul.opendev.org/t/openstack/build/d33fa43a546446639f9aa189e341d091/log/job-output.txt#11766 but then it ends up installing the default release (quincy) from the distro https://zuul.opendev.org/t/openstack/build/d33fa43a546446639f9aa189e341d091/log/job-output.txt#3622
15:53:37 melwitt actually it's the other way around, it installs the quincy packages before pulling the container image for reef (L3622 < L11766 ... sigh). I'll try to figure out how it ended up with quincy only, given this
15:54:17 frickler where do you see quincy? e.g. the mon log says reef https://224a5c0545942c9a5cc7-5b0493bbf4e0cfcdf1eae0328a979da8.ssl.cf2.rackcdn.com/884313/46/check/nova-ceph-multistore/d33fa43/controller/logs/ceph/ceph/43acd3ae-f7d3-456c-af2b-229fd530d8bc/ceph-mon.np0037692866_log.txt
15:54:40 frickler (also, where does is actually matter?)
15:54:47 frickler s/is/it/
15:56:18 melwitt I don't yet. I ran the same setup locally and after it was done I ran 'ceph version' and it was quincy. CEPH_RELEASE=reef should install reef, so for me it's a problem
15:57:23 melwitt I have to work backward from that to try to find how that happened. I'm hoping I'll be able to see it in the upstream log so I don't have to run it again locally
16:00:42 frickler might be a good idea to add the output of "ceph versions" at the end of the devstack setup
16:22:16 melwitt all of the ceph logs in the job are showing reef ...
16:24:29 melwitt so I dunno wtf happened in my local run and I have to dig into it. sorry for the noise ... 😖
16:36:00 gouthamr thanks frickler; ashrodri will do this
16:36:00 gouthamr > might be a good idea to add the output of "ceph versions" at the end of the devstack setup
#openstack-nova - 2024-06-13
05:46:33 ralonsoh stephenfin, hello! quick question: I was playing with devstack master but using branch=2024.1
05:46:43 ralonsoh and this came out: https://review.opendev.org/c/openstack/devstack/+/902758
05:46:54 ralonsoh because Nova patch is in master, nova-api was failing
05:47:59 ralonsoh is that ok? shouldn't devstack 2024.2 (in some months) be able to spawn a 2024.1 environment? just asking
06:24:04 frickler ralonsoh: no, devstack isn't supposed to be backwards compatible, you always need to use the matching stable branch for what you want to deploy
06:24:31 ralonsoh understood, so we are good then
06:26:14 sean-k-mooney master does sometimes work with older release but wihtout pinning all the versions in stack rc its proably not going to deploy what you think it is. so ya i always checkout the relevent branch before trying to use it to deploy something
06:27:49 sean-k-mooney just overridieng TARGET_BRANCH has never worked reliably so i dont even try since that is not tested
09:54:19 opendevreview Stephen Finucane proposed openstack/placement master: tox: Simplify functional testenv definitions https://review.opendev.org/c/openstack/placement/+/919578
09:54:20 opendevreview Stephen Finucane proposed openstack/placement master: pre-commit: Add sphinx-lint https://review.opendev.org/c/openstack/placement/+/919580
09:54:20 opendevreview Stephen Finucane proposed openstack/placement master: Integrate pre-commit https://review.opendev.org/c/openstack/placement/+/919579
10:47:22 opendevreview Merged openstack/osc-placement master: Replace simplejson by built-in json https://review.opendev.org/c/openstack/osc-placement/+/919999
13:12:27 ozzzo_work Can anyone help with my "openstack server list" question?
13:12:47 sean-k-mooney proably ask away
13:15:45 ozzzo_work I'm using this stupid web client so I have no history; I'll email it to work from my home computer
13:16:41 sean-k-mooney oh i see
13:16:44 sean-k-mooney openstack server list --limit -1
13:16:50 sean-k-mooney you want to list all vms
13:16:57 sean-k-mooney https://paste.openstack.org/show/bYNdWz8brEgB2ovSGiUF/
13:17:04 ozzzo_work yes
13:17:40 sean-k-mooney i think "openstack server list --all-projects" will list all vms if you jsut dont have limit unless that changed with the sdk and its alwasy using pagination now
13:18:03 ozzzo_work In 5.5.0 the limit is 1000
13:18:11 ozzzo_work let me test in 6.6.0 real quick
13:19:28 sean-k-mooney if you have more vms then the api limit
13:19:34 sean-k-mooney you need to use https://docs.openstack.org/python-openstackclient/latest/cli/command-objects/server.html#cmdoption-openstack-server-list-marker
13:19:44 sean-k-mooney passing the last server uuid as the marker
13:19:50 sean-k-mooney and make mulitple requests
13:20:17 sean-k-mooney but assuming you dont hit the api limit and if you dont pass --limit i would expelct it to list htem all
13:20:45 ozzzo_work we have >1000 VMs in some regions
13:20:50 sean-k-mooney i am not sure if osc handeles automatic pagination if you dont pass a limit
13:24:04 ozzzo_work Theoretically I could rewrite my scripts to use pagination, but "--limit -1" was nice because it was simple. Is there anything I can do to make it work in 6.6.0? I tried setting a different api version but that didn't seem to make a difference

Earlier   Later