| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2024-06-18 | |||
| 08:55:55 | bauzas | the job run failed because FORCE=yes wasn't around https://storage.bhs.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_a38/922140/3/experimental/nova-mtty/a383bd1/job-output.txt | |
| 08:57:24 | bauzas | oh, [[ True != \y\e\s ]] | |
| 08:57:30 | bauzas | hmmmm | |
| 09:03:18 | bauzas | sean-k-mooney: nevermind, I think I found the solution | |
| 09:03:53 | opendevreview | Sylvain Bauza proposed openstack/nova master: WIP : Enable an experimental job for testing mtty https://review.opendev.org/c/openstack/nova/+/922140 | |
| 09:03:54 | opendevreview | Sylvain Bauza proposed openstack/nova master: WIP : mtty: add devstack compilation for fedora builds https://review.opendev.org/c/openstack/nova/+/918420 | |
| 09:33:42 | frickler | bauzas: seems qa has finally found a volunteer to make devstack run on noble? ;) | |
| 09:35:27 | bauzas | frickler: yeah because I need to use a quite new libvirt version for testing mtty :) | |
| 09:46:50 | sean-k-mooney | bauzas: ya its proably not using the TrueOrFalse macro | |
| 09:47:52 | sean-k-mooney | frickler: i have been meanign to try that too. i spun up a ubuntu vm when the beta was released but i dont think i got around to running devstack on it | |
| 09:48:40 | sean-k-mooney | frickler: do we have images in ci yet. i was debating if i waould spend time on the dib patch or not | |
| 09:52:37 | bauzas | looks like using brackets with 'yes' worked | |
| 09:52:51 | bauzas | thanks codesearch | |
| 09:55:57 | frickler | sean-k-mooney: yes, images are there and I proposed to run py312 unittests on noble as next step, but I won't get to look into testing devstack myself https://review.opendev.org/c/openstack/openstack-zuul-jobs/+/922132 | |
| 09:58:57 | sean-k-mooney | frickler: ack ya i think that makes more sense then using pyenv on debian | |
| 09:59:59 | sean-k-mooney | frickler: for what its worth i have no issue with useing debian or debian testing in our ci i just dont think we need to be compling our onw python even if we are doing it once in the image build when there is py312 aviabel in 24.04 and centos 10 stream already | |
| 10:02:20 | frickler | sean-k-mooney: yes, using the default python version for each distro seems more sensible to me, too. but I haven't seen any work on centos 10 yet, is it out already? | |
| 10:02:39 | sean-k-mooney | yep its out for about a month | |
| 10:02:54 | sean-k-mooney | https://mirror.stream.centos.org/10-stream/ | |
| 10:03:54 | sean-k-mooney | the generic cloud image is here https://composes.stream.centos.org/stream-10/production/latest-CentOS-Stream/compose/BaseOS/x86_64/images/CentOS-Stream-GenericCloud-10-20240617.0.x86_64.qcow2 | |
| 10:04:11 | sean-k-mooney | frickler: i dont think there has been an offical release annouchment yet | |
| 10:05:22 | sean-k-mooney | im hoping we can get ubuntu 24.04 into ci this cycle and c10s next cycle as a testing runtime with the 2025.1 the last release with c9s | |
| 10:06:03 | sean-k-mooney | i.e 2025.1 will be the upgrade release for both ubuntu 22.04->24.04 and centos stream 9 -> 10 | |
| 10:07:28 | frickler | IMO you'd better get both working at least with devstack this cycle, then, so the TC can add them to the support list for 2025.1 from the start | |
| 10:08:29 | sean-k-mooney | that woudl be ideal yes | |
| 10:08:58 | sean-k-mooney | the supported runtimes have generally been based on aviablity not devstack support in the past | |
| 10:09:31 | sean-k-mooney | so both are valid for inclustion and should be added by default based on our disto upgrade policy | |
| 10:09:54 | sean-k-mooney | assumign c10s has its offial release annouchment before the start of the next cycle | |
| 10:10:10 | sean-k-mooney | practically speakign however yes we need devstack to work | |
| 10:10:24 | sean-k-mooney | the main challange is likely the changes to pip isntalling | |
| 10:10:38 | sean-k-mooney | ubuntu should be in a good place but c9s is still using a venv | |
| 10:10:46 | sean-k-mooney | so we need to fix that this cycle | |
| 10:10:58 | sean-k-mooney | *still not useing a venv | |
| 10:11:20 | sean-k-mooney | i have not chekced if c10s has blocked "sudo pip install" or not | |
| 10:11:58 | sean-k-mooney | but thats the direction most distor seam to be going so i would not be surpprised if they have, although fedora has not yet so im unsure | |
| 11:45:36 | opendevreview | Fabian Wiesel proposed openstack/nova-specs master: Lazy Metadata Loading in order to Reduce Server Load https://review.opendev.org/c/openstack/nova-specs/+/922201 | |
| 12:16:39 | opendevreview | Fabian Wiesel proposed openstack/nova-specs master: Lazy Metadata Loading in order to Reduce Server Load https://review.opendev.org/c/openstack/nova-specs/+/922201 | |
| 12:35:28 | opendevreview | Fabian Wiesel proposed openstack/nova-specs master: Lazy Metadata Loading in order to Reduce Server Load https://review.opendev.org/c/openstack/nova-specs/+/922201 | |
| 12:55:14 | opendevreview | Fabian Wiesel proposed openstack/nova-specs master: Lazy Metadata Loading in order to Reduce Server Load https://review.opendev.org/c/openstack/nova-specs/+/922201 | |
| 13:15:19 | bauzas | frickler: sean-k-mooney: fwiw, I got some distutils issue when stacking on noble https://zuul.opendev.org/t/openstack/build/9c2329f04a5344b4bbc9661e2360d282/log/job-output.txt#8152-8160 | |
| 13:16:27 | bauzas | thanks py3.12 | |
| 13:17:47 | sean-k-mooney | well distutils was remvoed | |
| 13:17:54 | sean-k-mooney | it wsa deprected for quite a while | |
| 13:18:10 | bauzas | I know | |
| 13:18:23 | bauzas | deprecated in 3.10 | |
| 13:18:26 | sean-k-mooney | so this seams to be a limitaiton in the requirement repo that should be updated | |
| 13:18:29 | bauzas | with a removal on 3.12 | |
| 13:18:37 | sean-k-mooney | its not an issue in openstack istslef | |
| 13:18:41 | bauzas | (which is *again* a short time) | |
| 13:18:52 | sean-k-mooney | 2 years is not short | |
| 13:19:01 | bauzas | for a language ? | |
| 13:19:06 | sean-k-mooney | correct | |
| 13:19:10 | bauzas | you're kidding? | |
| 13:19:13 | sean-k-mooney | i think your postion on that is unreasonable | |
| 13:19:21 | sean-k-mooney | but i dont want to get into that now | |
| 13:19:31 | bauzas | ask the operators what their positions are on something like 2 years | |
| 13:20:12 | bauzas | anyway, ship has sailed | |
| 13:20:22 | bauzas | and I agree, this is a requirements repo | |
| 13:20:22 | frickler | ah, nice, I noticed earlier today that also the use of pkg_resources in reqs repo needs replacing | |
| 13:20:29 | sean-k-mooney | its the cadance we do downtream major releease on and that lts distos often release on | |
| 13:21:05 | sean-k-mooney | also i belive as of 3.11 python are not giving 3 release or 3 years for removals | |
| 13:21:07 | frickler | also seems that there is no unit testing for edit-constraints. more fun | |
| 13:21:28 | sean-k-mooney | frickler: i would expect it to be used in all the tox gets for the cross gating | |
| 13:21:51 | sean-k-mooney | frickler:but probaly nothing direcltly testing it | |
| 13:22:12 | frickler | yes, the cross tests all run the latest support python, which is 3.11 | |
| 13:22:23 | sean-k-mooney | frickler: since this was not seen on the debinian pyevn tests im guessign it was falling back to the dist utils module form the default versio | |
| 13:23:08 | frickler | sean-k-mooney: iiuc the error is seen in devstack only, that didn't run on py312 yet | |
| 13:23:13 | bauzas | frickler: want me to report a bug against requirements ? | |
| 13:23:53 | sean-k-mooney | maybe we dont do the requriemts edit for tox because we are installing the package explcity | |
| 13:25:00 | sean-k-mooney | bauzas: can you try installing "sudo apt-get install python3-distutils" | |
| 13:25:16 | sean-k-mooney | bauzas: it might still be shipped but not instlled by default on 24.04 | |
| 13:25:21 | frickler | bauzas: I won't advise you to use storyboard and I very rarely look at it, but feel free to do so https://storyboard.openstack.org/#!/project/openstack/requirements | |
| 13:25:25 | bauzas | I don't have a box | |
| 13:25:31 | bauzas | I only use CI resources | |
| 13:25:47 | bauzas | but I can spin some noble instance (provided I can on our internal lab) | |
| 13:25:48 | frickler | I have a noble test node running, let me check that | |
| 13:26:00 | sean-k-mooney | oh ok thats fine ill see if the package exist and push a quick patch to devstack to add it | |
| 13:26:20 | bauzas | distutils is continued to be shipped thru setuptools | |
| 13:26:59 | bauzas | https://docs.python.org/3.12/whatsnew/3.12.html#distutils | |
| 13:27:10 | bauzas | so I see a workaround here | |
| 13:27:23 | sean-k-mooney | ok ya for the short term | |
| 13:27:50 | sean-k-mooney | ubuntu dropped the seperate pacakage in 23.10 | |
| 13:28:25 | sean-k-mooney | funny that https://packages.ubuntu.com/noble/python3-distutils-extra is still a thing | |
| 13:28:38 | sean-k-mooney | that pulls in setuptools to provide the core supprot | |
| 13:28:59 | bauzas | that's call "washing our hands" | |
| 13:29:03 | bauzas | called* | |
| 13:29:18 | sean-k-mooney | how, we are still going to need to fix requireemtns | |
| 13:29:25 | sean-k-mooney | this is just a short term workaround | |
| 13:29:27 | frickler | hmm, I don't get the failure on my noble node. and I didn't even run devstack there | |
| 13:29:40 | sean-k-mooney | do you have setuptools | |
| 13:29:52 | frickler | but it does have python3-dev installed, which pulled a lot of deps | |
| 13:30:04 | frickler | yes, python3-setuptools is one of them | |
| 13:30:17 | sean-k-mooney | ya so thats what provides it | |
| 13:30:35 | sean-k-mooney | we are surly installing python3-dev in devstack however | |
| 13:30:40 | sean-k-mooney | i wonder is the a venv thing | |
| 13:30:40 | bauzas | https://zuul.opendev.org/t/openstack/build/9c2329f04a5344b4bbc9661e2360d282/log/job-output.txt#8152-8160 | |
| 13:30:50 | bauzas | I can try to look at the installed packages | |
| 13:30:54 | sean-k-mooney | we dont allow site packages | |
| 13:31:09 | sean-k-mooney | so we might need to explcitly install setuptools form pip in the venv | |
| 13:31:12 | bauzas | https://0dcb2cbd4fa467880edd-6556216e0c08650db1b5109918fbb7f6.ssl.cf2.rackcdn.com/922140/4/experimental/nova-mtty/9c2329f/controller/logs/dpkg-l.txt | |