| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2023-05-12 | |||
| 08:32:38 | opendevreview | Danylo Vodopianov proposed openstack/nova master: Packed virtqueue support was added. https://review.opendev.org/c/openstack/nova/+/876075 | |
| 09:48:56 | opendevreview | Alexey Stupnikov proposed openstack/nova stable/xena: Remove deleted projects from flavor access list https://review.opendev.org/c/openstack/nova/+/883014 | |
| 09:52:29 | SvenKieske | hey there, we're currently implementing changes in kolla-ansible so nova uses service-tokens to talk to cinder, to address the vuln released 2 days ago. we hit an 500 Server Error during volume attachment. Should I report a bug, or might this just be a spurious failure? | |
| 09:53:05 | SvenKieske | our change: https://review.opendev.org/c/openstack/kolla-ansible/+/882893 the logged error: https://storage.bhs.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_11b/882893/3/check/kolla-ansible-rocky9-cephadm/11b1baf/primary/logs/ansible/test-core-openstack | |
| 09:54:16 | SvenKieske | Command is: "openstack server add volume kolla_boot_test test_volume --device /dev/vdb" | |
| 09:54:50 | SvenKieske | Error: HttpException: 500: Server Error for url: http://192.0.2.10:8774/v2.1/servers/2177019c-a915-4bda-a898-4fa806901144/os-volume_attachments, Unexpected API Error. | |
| 09:55:33 | opendevreview | Amit Uniyal proposed openstack/nova master: WIP: Delete dangling bdms https://review.opendev.org/c/openstack/nova/+/882284 | |
| 09:55:33 | opendevreview | Amit Uniyal proposed openstack/nova master: WIP: Reproducer for dangling volumes https://review.opendev.org/c/openstack/nova/+/881457 | |
| 10:53:04 | bauzas | SvenKieske: what's saying the nova-api log ? | |
| 10:54:04 | SvenKieske | bauzas: I'll take a look | |
| 10:54:26 | bauzas | 3b4c88ba497e9 a6ca4c77593c46a2b5e6c3c651e7e436 - - default default] HTTP exception thrown: Instance kolla_boot_test could not be found. | |
| 10:54:26 | bauzas | 2023-05-11 16:15:22.782 25 DEBUG nova.compute.api [None req-92ee72e6-a5e8-47b1-8428-47a71835f010 6ffc4ae1295442999283b4c88ba497e9 a6ca4c77593c46a2b5e6c3c651e7e436 - - default default] Failed to fetch instance by id kolla_boot_test get /var/lib/kolla/venv/lib64/python3.9/site-packages/nova/compute/api.py:2983 2023-05-11 16:15:22.782 25 INFO nova.api.openstack.wsgi [None req-92ee72e6-a5e8-47b1-8428-47a71835f010 6ffc4ae129544299928 | |
| 10:59:23 | bauzas | I don't see any instance creation in the nova-api.txt log | |
| 11:02:23 | bauzas | yeah, as well here https://storage.bhs.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_11b/882893/3/check/kolla-ansible-rocky9-cephadm/11b1baf/primary/logs/kolla/nova/nova-api-access.txt | |
| 11:04:09 | bauzas | and I don't see in the nova-api logs where os-volume_attachments is called :( | |
| 12:15:06 | SvenKieske | bauzas: I think I got an error/stacktrace, preparing a pastebin.. | |
| 12:22:07 | SvenKieske | bauzas: https://paste.opendev.org/show/819994/ | |
| 12:25:32 | SvenKieske | what's interesting is that it throws a 401 in the first few lines but what I get at the high level cli command is a 500 server error..mhm let me double check if this is really correlated | |
| 12:27:13 | bauzas | SvenKieske: yeah ok, so that's indeed due to https://review.opendev.org/c/openstack/kolla-ansible/+/882893 | |
| 12:29:29 | SvenKieske | yeah, but I see there are even more errors in keystone... it can't find the service project.. I'll circle back with my fellow kolla contributors who know more than me about our test setup, thanks so far! | |
| 12:33:28 | bauzas | ack | |
| 12:54:53 | opendevreview | Danylo Vodopianov proposed openstack/nova master: Packed virtqueue support was added. https://review.opendev.org/c/openstack/nova/+/876075 | |
| 13:19:23 | spatel | sean-k-mooney morning! I have question if you around | |
| 13:26:14 | bauzas | spatel: all RH folks (except me) are on PTO today | |
| 13:26:26 | bauzas | call me the highlander for today | |
| 13:27:23 | spatel | bauzas haha! Thanks.. I had question related hugepage. I have compute nodes with 256GB memory and does it worth enable hugepage on compute nodes? | |
| 13:27:44 | spatel | can i do memory overcommit with hugepage? | |
| 13:28:07 | bauzas | spatel: hugepages are set by using numa nodes | |
| 13:28:30 | bauzas | spatel: oh, you shouldn't (overcommit memory with hugepages) | |
| 13:28:58 | spatel | hmm! | |
| 13:29:03 | bauzas | like, playing with cpu pinning with overcommitted vCPUs (albeit this one, I'm sure it won't work) | |
| 13:29:52 | spatel | I don't want cpu-pinning. all i want no OOM so thinking to set hugepage | |
| 13:30:50 | bauzas | spatel: I have another internal prio now, so I can't really dig into code, but I think you need to respect MEMORY_MB with care | |
| 13:31:09 | spatel | bauzas okk | |
| 13:32:11 | spatel | Thank you bauzas | |
| 13:33:25 | bauzas | spatel: sorry to not able to help you more | |
| 13:33:40 | bauzas | spatel: but have you looked at our upstream docs for hugepages ? | |
| 13:34:27 | spatel | Yes! I am running hugepages on one of my cloud with cpu-pinning with numa scheduling | |
| 13:34:48 | spatel | This is new cloud where thinking to not use pinning but just hugepage | |
| 13:59:28 | sahid | sean-k-mooney: o/ regarding https://review.opendev.org/c/openstack/os-vif/+/881751, based on the bug report you wanted to backport it, right? | |
| 14:00:32 | bauzas | sahid: yes, sean-k-mooney in on PTO but yeah | |
| 14:00:33 | opendevreview | Jay Faulkner proposed openstack/nova-specs master: Re-Propose "Ironic Shards" for Bobcat/2023.2 https://review.opendev.org/c/openstack/nova-specs/+/881643 | |
| 14:02:01 | sahid | bauzas: ok, give me a moment I will cherry-pick them, thank you | |
| 14:02:46 | opendevreview | Sahid Orentino Ferdjaoui proposed openstack/os-vif stable/2023.1: set default qos policy https://review.opendev.org/c/openstack/os-vif/+/883015 | |
| 14:03:33 | opendevreview | Sahid Orentino Ferdjaoui proposed openstack/os-vif stable/zed: set default qos policy https://review.opendev.org/c/openstack/os-vif/+/883016 | |
| 14:03:36 | JayF | bauzas: sean-k-mooney: Updated the Ironic shards spec as requested, thank you! | |
| 14:03:55 | bauzas | JayF: fast-approved | |
| 14:04:42 | JayF | \o/ | |
| 14:18:02 | opendevreview | Merged openstack/nova-specs master: Re-Propose "Ironic Shards" for Bobcat/2023.2 https://review.opendev.org/c/openstack/nova-specs/+/881643 | |
| 14:21:51 | opendevreview | Sahid Orentino Ferdjaoui proposed openstack/os-vif stable/yoga: set default qos policy https://review.opendev.org/c/openstack/os-vif/+/883071 | |
| 14:25:17 | BobZAnnapolis__ | Hey folks, do you know of / have a documentatiuon location where i can find a detailed ERD for the NOVA dbs, the relations / fields defined, hopefully the purpose for each field, limitations, etc, something more than a "describe table" ? tia | |
| #openstack-nova - 2023-05-13 | |||
| 11:47:15 | opendevreview | Merged openstack/nova stable/zed: Use force=True for os-brick disconnect during delete https://review.opendev.org/c/openstack/nova/+/882860 | |
| 19:58:52 | opendevreview | Merged openstack/nova stable/zed: Enable use of service user token with admin context https://review.opendev.org/c/openstack/nova/+/882861 | |
| #openstack-nova - 2023-05-15 | |||
| 06:33:16 | opendevreview | Amit Uniyal proposed openstack/nova master: WIP: Delete dangling bdms https://review.opendev.org/c/openstack/nova/+/882284 | |
| 07:28:58 | dvo-plv | gibi, sean-k-mooney, sbauza: Hello. I have update nova code after os-traits release. Coud you please review packed ring support bp: https://review.opendev.org/q/bp/virtio | |
| 08:08:29 | bauzas | sean-k-mooney: could you please give a +2 on https://review.opendev.org/c/openstack/nova/+/882786 so I could prepare both a 2023.1 and Zed releases for both the CVEs ? | |
| 08:09:03 | bauzas | (or any stable core, like gibi) ^ | |
| 08:23:33 | opendevreview | Merged openstack/nova stable/2023.1: Bump nova-ceph-multstore timeout https://review.opendev.org/c/openstack/nova/+/882784 | |
| 09:03:48 | zigo | Does anyone know what I can do? | |
| 09:03:48 | zigo | This is in the new unit test test_cleanup_disconnect_volume() | |
| 09:03:48 | zigo | https://paste.opendev.org/show/br48JgltG6neK70WbkQA/ | |
| 09:03:48 | zigo | Hi! I'm trying to backport https://review.opendev.org/c/openstack/nova/+/882869 to Victoria. So far, it seems working, except, I have a unit test failure: | |
| 09:04:13 | zigo | I really would prefer to keep that unit test if possible ... | |
| 09:04:29 | zigo | melwitt: ^ | |
| 09:58:52 | dvo-plv | bauzas: Hello. I have update nova code after os-traits release. Coud you please review packed ring support bp: https://review.opendev.org/q/bp/virtio | |
| 11:18:49 | opendevreview | Amit Uniyal proposed openstack/nova-specs master: Adds cleanup to remove dangling volumes https://review.opendev.org/c/openstack/nova-specs/+/878757 | |
| 11:30:52 | sean-k-mooney | dvo-plv: the code looks ok but your missing test coverage | |
| 11:41:10 | dvo-plv | seean-k-mooney: We would like to implement tests after code approval in the separate bp. Does it possible, or we should to create a test immediately? | |
| 11:42:12 | sean-k-mooney | dvo-plv: no | |
| 11:42:23 | sean-k-mooney | we do not merge untested code in genral | |
| 11:42:40 | sean-k-mooney | so you need to have tests either in the same commmit or a followup commit | |
| 11:43:00 | sean-k-mooney | if its in a followup we will hold the first commit until we are happy with both and then merge them togeter | |
| 11:44:33 | sean-k-mooney | for openstack we treat master as if all commits could be deployed to production. | |
| 11:44:58 | sean-k-mooney | for lareger feature we genreally gate multi commit feature by having the change that enables the feature to be used last in the serise | |
| 11:45:37 | dvo-plv | I think we will implement and update this commit with tests coverage | |
| 11:45:49 | sean-k-mooney | that works | |
| 15:31:50 | opendevreview | Merged openstack/nova stable/yoga: Use force=True for os-brick disconnect during delete https://review.opendev.org/c/openstack/nova/+/882863 | |
| 15:33:45 | opendevreview | Merged openstack/nova stable/yoga: Enable use of service user token with admin context https://review.opendev.org/c/openstack/nova/+/882864 | |
| 17:24:47 | dansmith | fungi: if we change the wallaby implementation of the cve fix in gerrit (to address the thing raised in the bug), do we have to update errata or not? | |
| 17:24:56 | dansmith | the patch was in the bug, but wallaby should be out of bounds now | |
| 17:30:53 | fungi | technically we don't need to reference it because that branch is no longer under normal maintenance | |
| 17:31:27 | fungi | though i might amend the errata list with it and just not announce that so as to keep the noise level to a minimum | |
| 17:32:17 | fungi | we have a bit of flexibility with branches no longer under normal maintenance since those fixes are merely being included as a convenience and on a best-effort basis | |
| 17:32:27 | dansmith | ack, cool | |
| 17:32:50 | dansmith | I think you said "patches for older branches are appreciated but not required" so I just didn't know how that mapped to "provided as a courtesy and then changed" | |
| 17:32:52 | dansmith | melwitt: ^ | |
| 17:33:26 | fungi | it's sort of a grey area | |
| 17:33:40 | dansmith | I'm a walking grey area | |
| 17:33:49 | fungi | we didn't provide any non-nova wallaby patches anyway | |
| 17:34:14 | fungi | everything else stopped at xena (which is also no longer under normal maintenance at this point) | |
| 17:35:31 | melwitt | fungi: I had named the patch nova-2004555-xena_and_wallaby.patch erroneously thinking xena and wallaby were identical but it's really just nova-2004555-xena.patch | |
| 17:35:43 | fungi | makes sense | |
| 17:36:21 | fungi | for the downstream notification (and advisory of course) i split them up | |
| 17:36:37 | melwitt | oh, I see | |
| 17:37:33 | fungi | lots of the patches were "this applies to branches x and y" so i made separate identical branch x and branch y patches, because we don't want consumers to have to sort through a maze of patch name to branch mappings | |
| 17:41:07 | dansmith | melwitt: you're probably going to have to fix the job config on wallaby if you actually want to merge it too | |
| 17:41:32 | melwitt | fungi: that makes sense | |
| 17:41:41 | dansmith | the recent cinder tempest plugin changes for the ceph job have broken it.. I think just disabling the ceph job on wallaby is probably best | |
| 17:41:42 | melwitt | dansmith: good point | |
| 17:42:18 | melwitt | I at least wanted the current ps to be correct if anyone pulls it manually | |