| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2021-02-03 | |||
| 09:29:15 | kashyap | lyarwood: The tricky thing is that patch is a band-aid for some scenarios -- but will not fix the more difficult CPU-based live migration problems fully. | |
| 09:29:34 | kashyap | lyarwood: I've been thinking enough that I had a CPU model bug walk into my dream last night! | |
| 09:29:43 | lyarwood | haha | |
| 09:29:52 | kashyap | No kidding; I felt strange waking up | |
| 09:30:22 | kashyap | lyarwood: Given the bug OSP upgrades folks found ... for the fuller solution, a potential process trouble: | |
| 09:30:26 | lyarwood | kashyap: so that isn't going to resolve the the noTSX part? | |
| 09:31:13 | kashyap | lyarwood: It should, if your CPU problems are _just_ related to TSX | |
| 09:31:29 | lyarwood | kashyap: ack | |
| 09:31:53 | lyarwood | stephenfin: https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_f6e/769920/3/check/openstack-tox-py38/f6e5b16/testr_results.html btw, looks like my posion patch found *some* get_arch holes | |
| 09:31:54 | kashyap | lyarwood: The aforementioned "process trouble" is: I'm working w/ chengsheng (don't see them on IRC for a while) here (https://review.opendev.org/c/openstack/nova/+/762330/) to switch to newer CPU APIs | |
| 09:32:12 | stephenfin | lyarwood: oh boy :) | |
| 09:32:14 | kashyap | I've suggested a patch split of 5 pieces for ease of reviewing in my comment on PS12. | |
| 09:32:59 | stephenfin | lyarwood: Can you look at this. Should unblock circa 10 RBAC patches https://review.opendev.org/c/openstack/placement/+/771852/ | |
| 09:33:02 | stephenfin | *? | |
| 09:33:09 | lyarwood | stephenfin: ack | |
| 09:34:32 | kashyap | If chengsheng is busy; I'll split it (and give attribution where due, of course) as per the original spec ... which is not proposed for W -- that's the "too late to get in spec" process part; we uncovered this problem only recently. | |
| 09:34:38 | lyarwood | stephenfin: ah re the stable placement changes, I'm not a placement stable core | |
| 09:34:55 | stephenfin | oh, really? | |
| 09:35:11 | stephenfin | that sounds like another situation where we should be merging nova and placement | |
| 09:35:18 | stephenfin | gibi? ^ | |
| 09:35:25 | lyarwood | yeah it's a old list + stable-maint-core | |
| 09:35:31 | lyarwood | which I'm also not a member of | |
| 09:35:39 | gibi | hm | |
| 09:35:57 | gibi | do you suggest to add nova-stable-core to placement-stable-core? | |
| 09:35:59 | lyarwood | kashyap: yeah it's too late for W | |
| 09:36:10 | kashyap | lyarwood: Right, aware of that | |
| 09:36:18 | stephenfin | Yeah, I think that's a reasonable move | |
| 09:36:24 | lyarwood | kashyap: but we can still get it into a good state for X | |
| 09:36:31 | kashyap | Yep | |
| 09:36:38 | lyarwood | gibi: yes I think that would help | |
| 09:36:43 | kashyap | lyarwood: So, what I'm hoping is, if the band-aid backport above will fix most cases, we can get away with solid documentation of the issue. | |
| 09:36:56 | lyarwood | gibi: given we have outstanding .gitreview changes for stable/victoria still | |
| 09:36:57 | kashyap | (Which, I'm writing) | |
| 09:37:28 | lyarwood | kashyap: well downstream at least we could backport that | |
| 09:38:00 | kashyap | lyarwood: Sorry, backport what? The earlier-mentioned train backport? | |
| 09:38:13 | kashyap | If so, yes, of course | |
| 09:38:28 | lyarwood | the feature, it looks pretty self contained in the virt driver | |
| 09:38:38 | lyarwood | tbh you could even sell that as a bugfix in W | |
| 09:38:42 | lyarwood | for me at least | |
| 09:38:48 | lyarwood | if it's replacing broken behaviour | |
| 09:39:10 | lyarwood | stephenfin: ^ would you agree? | |
| 09:39:32 | lyarwood | stephenfin: context is https://review.opendev.org/c/openstack/nova/+/762330 | |
| 09:39:49 | stephenfin | This is about libvirt using the wrong APIs? | |
| 09:39:50 | lyarwood | stephenfin: switching over to a different libvirt API for checking CPUs | |
| 09:39:53 | lyarwood | yeah | |
| 09:40:06 | stephenfin | Yeah, I thought that was a bug tbh. Is someone saying otherwise? | |
| 09:40:15 | lyarwood | it's associated with a blueprint | |
| 09:40:17 | kashyap | lyarwood: Yeah, it could be argued as a bug-fix in good-faith | |
| 09:41:06 | lyarwood | cool, lets drop that reference and proceed as a bugfix | |
| 09:41:13 | kashyap | lyarwood: It is associated with a spec even: https://opendev.org/openstack/nova-specs/commit/70811da221035044e27 | |
| 09:41:35 | kashyap | See the work items for the recommended split of patches | |
| 09:42:04 | openstackgerrit | Wenping Song proposed openstack/nova master: Replaces tenant_id with project_id from List SG API https://review.opendev.org/c/openstack/nova/+/766726 | |
| 09:42:05 | openstackgerrit | Wenping Song proposed openstack/nova master: Replaces tenant_id with project_id from Flavor Access APIs https://review.opendev.org/c/openstack/nova/+/767704 | |
| 09:42:12 | lyarwood | MIN_{LIBVIRT,QEMU}_VERSION is beyond the required versions now so I'd say this is even more of a bugfix | |
| 09:42:33 | lyarwood | with some pre-written documentation we can use from the spec ;) | |
| 09:42:35 | kashyap | lyarwood: Yeap, yeap | |
| 09:44:13 | gibi | lyarwood, stephenfin: sent a mail about placement-stable-maint to ML http://lists.openstack.org/pipermail/openstack-discuss/2021-February/020186.html | |
| 09:44:23 | lyarwood | gibi: many thanks | |
| 09:44:39 | gibi | I will couple of days and then I will make the change | |
| 09:45:55 | lyarwood | ack | |
| 10:02:06 | elod | lyarwood: kashyap: +2+W'd >>> https://review.opendev.org/c/openstack/nova/+/758763 | |
| 10:02:16 | kashyap | elod: Thanks! | |
| 10:02:28 | elod | kashyap: np :) | |
| 11:36:08 | openstackgerrit | Lee Yarwood proposed openstack/nova master: fup: Merge duplicate volume attachment checks https://review.opendev.org/c/openstack/nova/+/773380 | |
| 13:45:52 | sean-k-mooney | V comes after U ... | |
| 13:46:11 | sean-k-mooney | meaning i need to backport to Victoria then Ussuri | |
| 13:51:24 | openstackgerrit | sean mooney proposed openstack/nova stable/victoria: only wait for plugtime events in pre-live-migration https://review.opendev.org/c/openstack/nova/+/770744 | |
| 13:56:28 | sean-k-mooney | lyarwood: can you take a look at ^ again | |
| 13:56:37 | sean-k-mooney | now that its in the correct order | |
| 14:01:14 | bauzas | sean-k-mooney: fwiw, I'm splitting my -W change for routed networks | |
| 14:01:27 | bauzas | the functests are saying OK so we know it works | |
| 14:01:42 | bauzas | (for all the move operations) | |
| 14:02:33 | sean-k-mooney | ok | |
| 14:03:06 | sean-k-mooney | i assume your going to either move teh func tests or the unit tests to a second patch? | |
| 14:03:47 | sean-k-mooney | for this type of change i trust the func tests more so would prefer those to stay with the code but ill leave that up to you | |
| 14:03:56 | sean-k-mooney | we will be reviewing the serise as a whole anyway | |
| 14:07:08 | bauzas | no, I'll have the functests in the same change for the filter | |
| 14:08:11 | bauzas | but I'll move the renaming of ensure_network_metadata() in a dependent change | |
| 14:10:02 | sean-k-mooney | cool | |
| 14:48:18 | gibi | fyi cores, I'm +2 on the vnc password patch https://review.opendev.org/c/openstack/nova/+/622336/ but it needs a second set of eyes | |
| 14:56:56 | sean-k-mooney | hum that touches a bunch of files but most are small changes it seams | |
| 14:57:13 | sean-k-mooney | im not a core but ill take a look at it | |
| 14:57:31 | gibi | sean-k-mooney: cool, thanks | |
| 14:57:34 | sean-k-mooney | it dose not look that difficult to reiew | |
| 15:31:54 | sean-k-mooney | gibi: -1 because i have a question on allowing console without a password to be created when the nova config has [vnc]/auth_schemes = none,vnc,vencrypt | |
| 15:32:03 | gibi | sean-k-mooney: thanks | |
| 15:32:07 | sean-k-mooney | gibi: but other then that the code looks ok | |
| 15:33:05 | sean-k-mooney | my comments are on the patch but while i woudl prefer not to save the pasword in plain text in novas db it is out of scope to use barbican so i view that as a future imporvment not a requirement for the feature | |
| 15:41:04 | openstackgerrit | Balazs Gibizer proposed openstack/nova stable/ussuri: Reproduce bug 1907522 in functional test https://review.opendev.org/c/openstack/nova/+/773930 | |
| 15:41:07 | openstack | bug 1907522 in OpenStack Compute (nova) victoria "test_*_with_qos_min_bw_allocation fails in the nova-multi-cell job with: nova.exception.MigrationPreCheckError: Migration pre-check error: Failed to create port bindings for host |
|
| 15:42:27 | openstackgerrit | Balazs Gibizer proposed openstack/nova stable/ussuri: Fallback to same-cell resize with qos ports https://review.opendev.org/c/openstack/nova/+/773932 | |
| 16:08:17 | melwitt | kashyap: took me awhile (sorry) but here's the qemu bug I opened about the block copy taking a long time to end the job after the last block is done https://bugs.launchpad.net/qemu/+bug/1914282 | |
| 16:08:18 | openstack | Launchpad bug 1914282 in QEMU "block copy job sometimes hangs on the last block for minutes" [Undecided,New] | |
| 16:08:34 | kashyap | melwitt: Hello | |
| 16:08:41 | melwitt | hello | |
| 16:09:19 | kashyap | melwitt: Thanks for filing; no need for sorry :-) | |
| 16:09:27 | kashyap | melwitt: Nice write-up | |
| 16:09:37 | kashyap | melwitt: One small comment; do you want to attach the pastebin content as a text file to the bug? | |
| 16:09:43 | kashyap | As pastebins expire | |
| 16:10:04 | melwitt | kashyap: oh, yeah, I will do that | |