Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2024-10-15
16:30:49 bauzas there will also be an i18n discussion I'll attend
16:31:07 bauzas I'll refer to it too with possibly follow-up discussions during nova sessions
16:31:36 bauzas #topic Stable Branches
16:31:42 bauzas elodilles: floor is yours
16:31:45 elodilles ack
16:31:49 elodilles #info stable/202*.* gates seem to be OK
16:32:03 elodilles #info final 2023.1 Antelope nova release before the branch transitions to 'unmaintained': https://review.opendev.org/c/openstack/releases/+/932406
16:32:08 elodilles does anyone have any bug fix that should be part of the final antelope release?
16:32:19 elodilles open & merged patches on stable/2023.1: https://etherpad.opendev.org/p/nova-stable-2023.1-eom
16:32:58 elodilles there are some patches that wait for a 2nd +2, but otherwise i don't know if anything would be important
16:33:13 elodilles and that's all from me
16:33:38 tkajinam maybe we should check if all follow-up fixes for image inspector have been backported
16:34:27 elodilles tkajinam: good point. afair, we had quite many of that backported,
16:34:44 elodilles though i don't know if that were all
16:35:05 bauzas well, there are a bunch of them
16:35:19 sean-k-mooney i think for the issues that are already fix the answer is yes
16:35:31 bauzas if you speak of the CVE fixes, yeah
16:35:41 tkajinam ah, ok
16:35:51 sean-k-mooney we also backported the fix for the aws and iso regressions on stable
16:35:52 elodilles https://review.opendev.org/q/topic:%22format-inspector%22+branch:stable/2023.1
16:35:57 bauzas if that's about the 'improvements' (like raw means raw), then I'm unsure
16:36:17 elodilles these are the ones that i'm aware of, at least (plus the bug fixes)
16:36:45 bauzas those are the CVE fixes
16:37:17 tkajinam My thought was mostly about regression fixes, not improvements. but I think we are ok about the known issues.
16:37:49 elodilles ACK
16:38:22 bauzas yeah so I checked
16:38:37 bauzas https://review.opendev.org/c/openstack/nova/+/926144 is more a signal patch, like 'dude, you shouldn't be doing that'
16:39:35 bauzas and this was following https://review.opendev.org/c/openstack/nova/+/924866 which was backported
16:39:56 tkajinam lgtm
16:40:31 bauzas we recently merged https://review.opendev.org/c/openstack/nova/+/930754 but I don't see any interest in backporting it down
16:40:51 bauzas and there is a dep on glance
16:41:08 bauzas so I think we can't use the glance safebelt
16:41:17 bauzas hence the non-backport
16:41:30 bauzas so I guess we're cool
16:42:02 bauzas moving on ten
16:42:06 bauzas then*
16:42:09 elodilles ++
16:42:23 bauzas #topic vmwareapi 3rd-party CI efforts Highlights
16:42:31 bauzas fwiesel: anything to mention ?
16:42:43 fwiesel No, nothing from my side.
16:42:48 fwiesel Sorry for the lack of progress.
16:43:29 bauzas no worries, we're all busy
16:43:35 bauzas particularly me.
16:43:51 bauzas #topic Open discussion
16:43:57 bauzas (tkajinam): Removing db migration tool from nova_placement to placement: https://review.opendev.org/c/openstack/placement/+/932324
16:43:59 bauzas shoot
16:44:28 tkajinam I discussed this already with sean-k-mooney and gibi but I think it may make sense to bring the topic here to get any additional feedback
16:45:10 tkajinam placement repository has carried the migration tool which pulls placement db from nova db. This was required in upgrade from Stein to Train and I'm wondering if we can directly drop it now because I don't think anyone may use it from master nowadays
16:45:45 bauzas this is old code
16:45:55 bauzas I don't see any controversy in removing it
16:46:09 tkajinam :-)
16:46:15 bauzas and people wanting to use it could just pip install in a venv an older release of placement
16:46:23 bauzas and use that tool
16:46:31 gibi yeah, drop it.
16:46:50 tkajinam A quick question I still have is whether we should keep the upgrade note which is almost empty https://8f5091142d0d93e54bb3-66f76d6fb4c84b410723fddf17d0dbe7.ssl.cf5.rackcdn.com/932324/2/check/openstack-tox-docs/93314e0/docs/admin/upgrade-notes.html or keep it (and the administrator guide section) as a placeholder
16:47:03 opendevreview Amit Uniyal proposed openstack/nova master: Update Nova bdm with updated swap info https://review.opendev.org/c/openstack/nova/+/929858
16:47:04 bauzas from a procedural perspective, I litterally see zero paperwork to fill
16:48:20 bauzas tkajinam: keep that
16:48:26 gibi tkajinam: that upgrade-notes seem still relevant
16:48:35 bauzas running a status check pre-flight is important
16:48:50 tkajinam yes
16:49:10 bauzas anyway, I think we're done with that item, we'll review it
16:49:20 tkajinam ok then https://review.opendev.org/c/openstack/placement/+/932324 is ready. I had to wipe previous +2 to fix doc issues
16:49:22 tkajinam thanks !
16:49:23 bauzas tkajinam: feel free to add your patch into the status etherpad
16:49:42 tkajinam will do
16:49:48 bauzas ++
16:50:06 bauzas another item, less easy to answer
16:50:08 bauzas (tkajinam): Can we remove (or at least deprecate) WIndows OS support ?
16:50:21 bauzas any good reason for doing it ?
16:50:55 bauzas oh, the hypervisor bits
16:51:11 tkajinam So as you know WinStackers project has been retired and now OpenStack on Windows is no longer maintained
16:51:12 bauzas for a second, I thought it was about the image type and the instances :)
16:51:41 tkajinam I've seen some issues with cross platform support when working on replacing unmaintained libs (eg netifaces)
16:52:16 tkajinam I wonder if we can completely drop the windows support if it's not really tested/maintained or at least deprecate it so that we can get rid of it when something is completely broken
16:52:23 bauzas if that's removing tech debt on things that we removed but left remains, sure I second your work
16:53:14 bauzas anyone having any concerns about that ?
16:54:11 bauzas I checked and apparently you can install a libvirt client on windows
16:54:39 bauzas it would require your hypervisor to be remote tho, which is something we don't support
16:54:48 tkajinam yes
16:54:54 tkajinam as we don't support hyper-v via libvirt
16:55:02 tkajinam I personally prefer direct removal given the fact hyper-v driver was already removed, but if we want to do it safe then we can deprecate it this cycle. We can discuss the strategy in the review https://review.opendev.org/c/openstack/nova/+/932407
16:55:29 sean-k-mooney ya so hyperv via libvirt ws never supported
16:55:39 bauzas I just think we can process with reviewing your patch
16:55:49 bauzas nothing controversial so far
16:55:53 tkajinam thanks
16:56:27 bauzas anything else before we wrap up ?
16:56:53 tkajinam just fyi. I've added a few bugs (and their fixes) to the bottom of status etherpad https://etherpad.opendev.org/p/nova-2025.1-status (L87-91)
16:57:14 sean-k-mooney tkajinam: we have some windows supprot we can delete in os-vif too by the way
16:57:15 tkajinam these replaces the deprecated/unmaintained libs so I appreciate reviews for these.
16:57:28 tkajinam sean-k-mooney, yeah I remember I did deprecate these in the past cycle
16:58:01 bauzas tkajinam: ack
16:58:07 tkajinam so it maybe time to drop it, as we usually use eventlet patch differently for windows removing the logic may help removing eventlet.
16:58:27 gmann tkajinam: what exactly left to remove window support?
16:58:41 tkajinam gmann, tons of os.name == 'nt' checks
16:58:42 bauzas gmann: see above, there is a cleanup patch
16:58:59 tkajinam maybe not 'tons' but 'numbers'
16:59:05 gmann ohk, just cleanup things. got it
16:59:28 gmann but we have already declared the winbdow support removal during hyper-V removal
16:59:49 tkajinam yes

Earlier   Later