January 2023 - openstack-discuss - lists.openstack.org

Experience with VGPUs
by Oliver Weinmann 22 Jun '23

22 Jun '23

Dear All, we are planning to have a POC on VGPUs in our Openstack cluster. Therefore I have a few questions and generally wanted to ask how well VGPUs are supported in Openstack. The docs, in particular: https://docs.openstack.org/nova/zed/admin/virtual-gpu.html explain quite well the general implementation. But I am more interested in general experience with using VGPUs in Openstack. We currently have a small YOGA cluster, planning to upgrade to Zed soon, with a couple of compute nodes. Currently our users use consumer cards like RTX 3050/3060 on their laptops and the idea would be to provide VGPUs to these users. For this I would like to make a very small POC where we first equip one compute node with an Nvidia GPU. Gladly also a few tips on which card would be a good starting point are highly appreciated. I know this heavily depends on the server hardware but this is something I can figure out later. Also do we need additional software licenses to run this? I saw this very nice presentation from CERN on VGPUs: https://indico.cern.ch/event/776411/contributions/3345183/attachments/18516… In the table they are listing Quadro vDWS licenses. I assume we need these in order to use the cards? Also do we need something like Cyborg for this or is VGPU fully implemented in Nova? Best Regards, Oliver

17 36

[ironic][stable] Proposing EOL of ironic project branches older than Wallaby
by Jay Faulkner 01 Jun '23

01 Jun '23

Hi all, Ironic has a large amount of stable branches still in EM. We need to take action to ensure those branches are either retired or have CI repaired to the point of being usable. Specifically, I'm looking at these branches across all Ironic projects: - stable/queens - stable/rocky - stable/stein - stable/train - stable/ussuri - stable/victoria In lieu of any volunteers to maintain the CI, my recommendation for all the branches listed above is that they be marked EOL. If someone wants to volunteer to maintain CI for those branches, they can propose one of the below paths be taken instead: 1 - Someone volunteers to maintain these branches, and also report the status of CI of these older branches periodically on the Ironic whiteboard and in Ironic meetings. If you feel strongly that one of these branches needs to continue to be in service; volunteering in this way is how to save them. 2 - We seriously reduce CI. Basically removing all tempest tests to ensure that CI remains reliable and able to merge emergency or security fixes when needed. In some cases; this still requires CI fixes as some older inspector branches are failing *installing packages* in unit tests. I would still like, in this case, that someone volunteers to ensure the minimalist CI remains happy. My intention is to let this message serve as notice and a waiting period; and if I've not heard any response here or in Monday's Ironic meeting (in 6 days), I will begin taking action on retiring these branches. This is simply a start; other branches (including bugfix branches) are also in bad shape in CI, but getting these retired will significantly reduce the surface area of projects and branches to evaluate. I know it's painful to drop support for these branches; but we've provided good EM support for these branches for a long time and by pruning them away, we'll be able to save time to dedicate to other items. Thanks, Jay Faulkner

3 5

[Openstack] Lack of Balance solution such as Watcher.
by Nguyễn Hữu Khôi 16 Mar '23

16 Mar '23

I am building a private cloud.Everything is ok. But I cannot find a way to balance vm when It have heavy load which cause impact other vm on the compute node. I found solutions such as Watcher and Leveller. But they need to be done manually. Watcher is not good because It need cpu metric such as cpu load in Ceilometer which is removed so we cannot use it. Leveller is good, but It is obsolete. Nguyen Huu Khoi

4 7

live resize
by Parsa Aminian 15 Mar '23

15 Mar '23

hello on openstack with ceph backend is it possible to live resize instances ? I want to change flavor without any down time .

5 6

[security-sig] Polls in preparation to revive our meetings
by Jeremy Stanley 01 Mar '23

01 Mar '23

As discussed at the last PTG, the present meeting time (15:00 UTC on the first Thursday of each month) is inconvenient for some attendees, and that combined with year-end holidays and general busy weeks recently have led to skipping them entirely. In order to start narrowing down the potential meeting schedule, I have two initial polls. The first is to determine what frequency we should meet. If you have an opinion on that, please fill out this poll before Thursday, February 9 (two weeks from today): https://framadate.org/wz7GioqmgyWeILkr The second poll is to hopefully determine what day of the week is optimal for potential attendees. If you have a preference for which day of the week to meet, please complete this one by the same date as the first: https://framadate.org/CyxKgZPT8PWxcCnJ Once I can analyze the results, I'll put together a more specific poll for choosing a time of day as well as possibly choosing which week(s) of the month (if we don't settle on weekly frequency). In the meantime, let's plan to hold February's meeting on Thursday the 2nd at 15:00-16:00 UTC as usual[*] for anyone who is able to attend, and I'll get an agenda together in preparation for that. [*] https://meetings.opendev.org/#OpenStack_Security_SIG_meeting -- Jeremy Stanley

1 1

[puppet] Retiring untested and unmaintained modules (murano, rally and tacker)
by Takashi Kajinami 27 Feb '23

27 Feb '23

Hello, In Puppet OpenStack projects we have multiple modules to support multiple OpenStack components. However unfortunately some of these have not been attracting enough interest from developers and have been unmaintained. During the past few cycles we retired a few incomplete modules but I'm wondering if we can retire a few unmaintained modules now, to reduce our maintenance/release effort. I checked the modules we have currently, and I think the following three can be first candidates. - puppet-murano - puppet-rally - puppet-tacker We haven't seen any feedback from users about these modules for a long time. Most of the changes for the past 2~3 years are proposed by me but I am not really using these components. These modules do not have proper test coverage and it's quite difficult for us to catch any breakage and honestly I'm not quite sure these modules can work properly with the latest code. Actually we've often caught up with the latest requirements several years after the change was made in the software side, and I'm afraid these are not well-maintained. eg. - support for tacker-conductor was added 4 years after the service was added - we didn't noticed that the openstack plugin was split out from the core rally package for several years If anybody has concerns with retiring these modules, then please let us know. If we don't hear any objections for a while, then I'll start proposing changes for project retirement. Thank you, Takashi -- ---------- Takashi Kajinami

4 9

[kolla-ansible] [yoga] [magnum] [k8s] cannot attach persistent volume to pod
by Stuart Whitman 24 Feb '23

24 Feb '23

Hello, When I try to attach a persistent cinder volume to a pod, I get FailedMount and FailedAttachVolume timeout events. I also get these errors in the log of the csi-cinder-controllerplugin-0 pod: E1020 13:38:41.747511 1 reflector.go:126] k8s.io/client-go/informers/factory.go:133: Failed to list *v1beta1.VolumeAttachment: the server could not find the requested resource E1020 13:38:41.748187 1 reflector.go:126] k8s.io/client-go/informers/factory.go:133: Failed to list *v1beta1.CSINode: the server could not find the requested resource I fixed a CrashLoopBackoff error with the csi-snapshotter container in the csi-cinder-controllerplugin-0 pod by providing the label "csi_snapshotter_tag=v4.0.0" when I created the cluster template. I found that suggestion in an issue on the GitHub cloud-provider-openstack project. I'm not finding any help with this error on Google. Thanks, -Stu _____________________________________ The information contained in this e-mail and any attachments from Group W may contain confidential and/or proprietary information and is intended only for the named recipient to whom it was originally addressed. If you are not the intended recipient, be aware that any disclosure, distribution, or copying of this e-mail or its attachments is strictly prohibited. If you have received this e-mail in error, please notify the sender immediately of that fact by return e-mail and permanently delete the e-mail and any attachments to it.

3 5

[Magnum]enable cluster user trust
by Nguyễn Hữu Khôi 16 Feb '23

16 Feb '23

Hello guys. I am going to use Magnum for production but I see that https://nvd.nist.gov/vuln/detail/CVE-2016-7404 if I want to use cinder for k8s cluster. Is there any way to fix or minimize this problem? Thanks. Nguyen Huu Khoi

2 3

[heat][release] Proposing to EOL Train, Ussuri and Victoria
by Brendan Shephard 10 Feb '23

10 Feb '23

Hi folks, We’re looking to move Train, Ussuri and Victoria branches to EOL for Heat projects. Just reaching out to see if there are any objections, if I don’t hear anything back I’ll move forward with that this week. Cheers, Brendan Shephard Senior Software Engineer Red Hat Australia

2 3

[magnum] Other Distro for k8s
by Nguyễn Hữu Khôi 10 Feb '23

10 Feb '23

Hello guys. I know that Magnum is using Fedora Coreos for k8s. Why don't we use a long-term distro such as Ubuntu for this project? I will be more stable. and this project seems obsolete with the old version for k8s. Nguyen Huu Khoi

3 6