| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2021-08-09 | |||
| 09:07:07 | stephenfin | lyarwood: Any chance you could cast an eye over a little more of the DB series this week, in the absence of gibi and bauzas? It starts on this patch which _looks_ huge but was mostly an invocation of sed :) https://review.opendev.org/c/openstack/nova/+/799524/ | |
| 09:07:58 | lyarwood | yarp, was up most the night but II'll make a start once this coffee kicks in | |
| 09:08:03 | lyarwood | I'll* | |
| 09:38:38 | opendevreview | Stephen Finucane proposed openstack/nova-specs master: Amend configurable-instance-hostnames for more response changes https://review.opendev.org/c/openstack/nova-specs/+/803892 | |
| 09:49:10 | kashyap | stephenfin: Hey, when you get a min, maybe you can settle this dispute here: https://review.opendev.org/c/openstack/nova/+/798680/3/nova/virt/libvirt/config.py#2058 | |
| 09:51:15 | stephenfin | kashyap: I think you're correct. Those classes are used to build XML for new instances. I don't think we use it to parse anything for existing instances. | |
| 09:51:51 | kashyap | stephenfin: Exactly so; thank you very much. I was wondering if I was confusing myself | |
| 09:52:32 | stephenfin | unrelated: have we agreed on virtio? Doesn't that have impacts for Windows guests? | |
| 09:53:27 | kashyap | stephenfin: Right; I've discussed this at length w/ the QEMU graphics maint, and 'virtio' seems good for both | |
| 09:53:30 | kashyap | stephenfin: What more: | |
| 09:53:46 | kashyap | stephenfin: I got a Red Hat virt QE to test both Windows and Linux guests with 'cirrus' and do some live migration tests | |
| 09:54:15 | kashyap | (with 'cirrus', changed to 'virtio', i.e.) And the tests succeed. See my comment on PS 3 (Aug 04 4:27 PM) | |
| 09:54:21 | stephenfin | kashyap: What are your thoughts on whipping up a doc with the current and future default device models, the reasons for these choices, and the commands one can use to migrate existing instances from one to the other | |
| 09:54:45 | stephenfin | in lieu of a spec or anything else | |
| 09:55:15 | kashyap | stephenfin: "Device models" --> You're talking only of video, or others too? | |
| 09:55:22 | stephenfin | there's a lot of context embedded in various reviews, IRC convos and emails now. It would be good to get it all transcribed somewhere | |
| 09:55:40 | stephenfin | That depends. Are you personally just targeting video right now? | |
| 09:56:09 | kashyap | stephenfin: Yes, right now only 'video'. And yes, I fully agree: the context is across sources, and only those who're immersed in it can have a semblance of clarity. | |
| 09:56:19 | stephenfin | If so, yes, just video for now, but lets name the document generically (doc/source/user/libvirt-device-models.rst ?) so we can expand on it later | |
| 09:56:26 | stephenfin | presumably via lyarwood's efforts | |
| 09:56:27 | kashyap | stephenfin: I'm always up for docs such as what you describe. | |
| 09:57:00 | stephenfin | spot on. I'd be happy to review that. We can use it as a stand-in for a spec and point to it from other reviews | |
| 09:57:17 | stephenfin | Plus, we've decent docs afterwards | |
| 09:58:02 | kashyap | stephenfin: Yeah; I'll throw in a place-holder, so I that I won't forget it after PTO starting this Thu | |
| 09:58:31 | kashyap | stephenfin: Thanks, as always for your relentless care for thoughtful docs! :) | |
| 10:01:00 | kashyap | (Just to wrap-up the video model one temporarily: the move from 'cirrus' to 'virtio' is far less disruptive based on the tests we've done so far - due to the graceful fallback that 'virtio' provides in case there are guest drivers.) | |
| 10:01:57 | opendevreview | Stephen Finucane proposed openstack/nova-specs master: Revert "Amend configurable-instance-hostnames to include response changes" https://review.opendev.org/c/openstack/nova-specs/+/803895 | |
| 10:03:27 | stephenfin | gmann: I've proposed two competing spec amendments: one proposes removing all extension prefixes for various '/servers' and '/flavors' APIs, and the other proposes not removing any (so 'OS-EXT-SRV-ATTR:hostname' stays the same) | |
| 10:04:06 | stephenfin | gmann: The perfectionist in me would prefer the former. The realist says the latter might be wiser, at least for this microversion. | |
| 10:04:30 | stephenfin | gmann: https://review.opendev.org/c/openstack/nova-specs/+/803892 (remove all prefixes) https://review.opendev.org/c/openstack/nova-specs/+/803895 (remove no prefixes) | |
| 13:04:41 | slaweq | hi nova-stable-maint team, can You take a look at https://review.opendev.org/c/openstack/nova/+/791420 ? | |
| 13:05:07 | slaweq | it already have +2 from elodilles_pto but it needs second one :) | |
| 13:06:17 | lyarwood | ack will do | |
| 13:54:13 | gmann | stephenfin: ack, will check and comment, thanks | |
| 14:04:18 | lyarwood | stephenfin: https://zuul.opendev.org/t/openstack/build/6cedd9d4547c4f608553ace4701d4709 - did you see that docs failure on your https://review.opendev.org/c/openstack/nova/+/799523/ change btw? | |
| 14:21:37 | stephenfin | lyarwood: No, I did not | |
| 14:32:04 | stephenfin | lyarwood: Sigh, that failed 8 times in a row. I missed it because there were quite a few random failures elsewhere and I got lazy and stopped checking the logs :-( | |
| 14:32:18 | stephenfin | lyarwood: The issue is that this change should not be here but rather in the next patch in the series https://review.opendev.org/c/openstack/nova/+/799523/1/etc/nova/nova-config-generator.conf | |
| 14:33:00 | stephenfin | afraid so | |
| 14:33:07 | lyarwood | RIP the gate | |
| 14:33:21 | lyarwood | but cool at least it's fixed then | |
| 14:33:22 | stephenfin | Much maligned but ultimately missed | |
| 14:34:02 | lyarwood | the series looks good overall btw, I went through to the end but wanted to look at a few things in more detail before voting either way | |
| 14:34:30 | lyarwood | mostly with the automatic migrations at the end tbh | |
| 14:35:23 | stephenfin | anything else obvious you want fixed before I push this rebase? | |
| 14:35:40 | lyarwood | nothing at the moment no | |
| 14:35:42 | stephenfin | I dragged the follow-up into the series after the patch it fixed | |
| 14:35:47 | lyarwood | cool | |
| 14:38:06 | opendevreview | Stephen Finucane proposed openstack/nova master: db: Unify 'nova.db.api', 'nova.db.sqlalchemy.api' https://review.opendev.org/c/openstack/nova/+/799524 | |
| 14:38:06 | opendevreview | Stephen Finucane proposed openstack/nova master: db: Register database config options ourselves https://review.opendev.org/c/openstack/nova/+/799523 | |
| 14:38:07 | opendevreview | Stephen Finucane proposed openstack/nova master: db: Move remaining 'nova.db.sqlalchemy' modules https://review.opendev.org/c/openstack/nova/+/799525 | |
| 14:38:07 | opendevreview | Stephen Finucane proposed openstack/nova master: db: Exclude the 'use_db_reconnect' option https://review.opendev.org/c/openstack/nova/+/803778 | |
| 14:38:08 | opendevreview | Stephen Finucane proposed openstack/nova master: db: Avoid use of ALTER in initial migration https://review.opendev.org/c/openstack/nova/+/800076 | |
| 14:38:08 | opendevreview | Stephen Finucane proposed openstack/nova master: db: Post reshuffle cleanup https://review.opendev.org/c/openstack/nova/+/799526 | |
| 14:38:09 | opendevreview | Stephen Finucane proposed openstack/nova master: db: Add initial alembic migration for API DB https://review.opendev.org/c/openstack/nova/+/799528 | |
| 14:38:09 | opendevreview | Stephen Finucane proposed openstack/nova master: db: Add initial alembic migration for main DB https://review.opendev.org/c/openstack/nova/+/799527 | |
| 14:38:10 | opendevreview | Stephen Finucane proposed openstack/nova master: db: Trivial style changes https://review.opendev.org/c/openstack/nova/+/799529 | |
| 14:38:11 | opendevreview | Stephen Finucane proposed openstack/nova master: db: Integrate alembic https://review.opendev.org/c/openstack/nova/+/799530 | |
| 14:38:11 | opendevreview | Stephen Finucane proposed openstack/nova master: db: Normalize migrations tests https://review.opendev.org/c/openstack/nova/+/799684 | |
| 14:38:13 | opendevreview | Stephen Finucane proposed openstack/nova master: docs: Add documentation on database migrations https://review.opendev.org/c/openstack/nova/+/800078 | |
| 14:38:13 | opendevreview | Stephen Finucane proposed openstack/nova master: db: Enable auto-generation of migrations https://review.opendev.org/c/openstack/nova/+/800077 | |
| 14:38:15 | opendevreview | Stephen Finucane proposed openstack/nova master: db: Final cleanups https://review.opendev.org/c/openstack/nova/+/800484 | |
| 14:39:41 | stephenfin | okay, https://review.opendev.org/c/openstack/nova/+/799523/ now excludes the one line change to etc/nova/nova-config-generator.conf. https://review.opendev.org/c/openstack/nova/+/799524 is the opposite, in that it now includes this. Both otherwise unchanged | |
| 15:04:50 | opendevreview | Merged openstack/nova stable/rocky: [neutron] Get only ID and name of the SGs from Neutron https://review.opendev.org/c/openstack/nova/+/791420 | |
| 15:05:36 | opendevreview | Slawek Kaplonski proposed openstack/nova stable/queens: [neutron] Get only ID and name of the SGs from Neutron https://review.opendev.org/c/openstack/nova/+/791421 | |
| 16:55:22 | lyarwood | kashyap: https://zuul.opendev.org/t/openstack/build/3c4856a7e156479e8b7bda3135389b64/log/controller/logs/screen-n-cpu.txt?severity=4 - one for tomorrow but I'm seeing `qemu unexpectedly closed the monitor: Could not allocate dynamic translator buffer` errors in the CentOS job, grepping around I think this is SELinux related but could use a hand tracking it down. | |
| 18:19:44 | opendevreview | Merged openstack/nova master: db: Register database config options ourselves https://review.opendev.org/c/openstack/nova/+/799523 | |
| #openstack-nova - 2021-08-10 | |||
| 00:03:00 | opendevreview | Brin Zhang proposed openstack/nova master: Replaces tenant_id with project_id from Rebuild Server API https://review.opendev.org/c/openstack/nova/+/766380 | |
| 00:04:07 | opendevreview | Brin Zhang proposed openstack/nova master: Replaces tenant_id with project_id from List SG API https://review.opendev.org/c/openstack/nova/+/766726 | |
| 00:52:58 | opendevreview | Brin Zhang proposed openstack/nova master: Replaces tenant_id with project_id from Flavor Access APIs https://review.opendev.org/c/openstack/nova/+/767704 | |
| 01:08:17 | opendevreview | Brin Zhang proposed openstack/nova master: Replaces tenant_id with project_id from List/Show usage APIs https://review.opendev.org/c/openstack/nova/+/768509 | |
| 02:02:26 | opendevreview | Brin Zhang proposed openstack/nova master: Replace tenants* with projects* of policies https://review.opendev.org/c/openstack/nova/+/765315 | |
| 02:02:27 | opendevreview | Brin Zhang proposed openstack/nova master: Replace tenant_id with project_id in os-quota-sets path https://review.opendev.org/c/openstack/nova/+/768851 | |
| 02:02:27 | opendevreview | Brin Zhang proposed openstack/nova master: Replace os-simple-tenant-usage with os-simple-project-usage https://review.opendev.org/c/openstack/nova/+/768852 | |
| 02:02:28 | opendevreview | Brin Zhang proposed openstack/nova master: Replace tenant* with project* in codes https://review.opendev.org/c/openstack/nova/+/769329 | |
| 02:02:28 | opendevreview | Brin Zhang proposed openstack/nova master: Replace tenant_id with project_id in Limits API https://review.opendev.org/c/openstack/nova/+/768862 | |
| 02:09:16 | opendevreview | Takashi Kajinami proposed openstack/nova master: tests: Add test for bug #1936278 https://review.opendev.org/c/openstack/nova/+/802893 | |
| 07:01:29 | opendevreview | Slawek Kaplonski proposed openstack/nova stable/queens: [neutron] Get only ID and name of the SGs from Neutron https://review.opendev.org/c/openstack/nova/+/791421 | |
| 08:36:08 | kashyap | lyarwood: Looking | |
| 08:54:08 | kashyap | lyarwood: Okay, I see that error was already discussed upstream QEMU by Debian folks last year ... looking | |
| 09:05:47 | kashyap | lyarwood: Okay, talking to upstream QEMU maint: | |
| 09:06:38 | kashyap | - That terrible error is thrown becuase QEMU is unable execute code from the memory QEMU is executing from — IOW, it requires the allocated chunk of memory to be executable. | |
| 09:07:36 | kashyap | - Peter (from QEMU) had three possible guesses off-hand: | |
| 09:07:47 | kashyap | (a) host OS is enforcing no-rwx like OpenBSD can, or (b) if the user's set a very aggressive 'ulimit'on virtual memory usage, or (c) perhaps a system with no swap and not enough RAM either | |
| 09:11:33 | kashyap | - From QEMU 6.0 or higher, they fixed that usless error to report a better error including the error code: https://git.qemu.org/?p=qemu.git;a=commitdiff;h=6bc144237a857b | |
| 09:11:51 | kashyap | So, our possible options are: | |
| 09:12:38 | kashyap | (1) Check if this is repetable at all -- and rule out 'ulimit' and "not enough RAM" options | |
| 09:12:58 | kashyap | (2) Try QEMU 6.0 or higher, to identify the precise error - that may give better info to the QEMU folks to debug further | |
| 09:13:57 | kashyap | (3) In case we have access to the host, get `strace` output, which should tell us where the mmap() call is failing. | |
| 09:15:30 | kashyap | I don't think (2) and (3) are not possible because our hands are tied, and no access to the machine? :-( | |
| 09:16:32 | kashyap | Oh, correction: Peter tells me that the 'strace' will give the same info as trying with QEMU 6.0 or higher. So we can try whatever is eaiser: option two or (2) or (3) from the above. | |
| 09:19:50 | kashyap | s/two or// | |
| 09:27:05 | lyarwood | kashyap: kk, we don't have virt-preview for centos stream so I'm not sure what we can do tbh | |
| 09:27:20 | lyarwood | oh wait we do | |
| 09:27:26 | lyarwood | TIL | |
| 09:27:35 | lyarwood | cool I'll enable that now | |
| 09:27:40 | kashyap | lyarwood: Yeah, CentOS does have it, last I checked | |
| 09:27:49 | kashyap | lyarwood: Via a CI job? | |
| 09:27:59 | lyarwood | yeah | |
| 09:30:23 | kashyap | lyarwood: Ah, cool. In case this is not resolved by tomm afternoon (I'll be on PTO from day after), you can follow-up with 'pm215' (Peter Maydell) on #qemu, OFTC. He helped the initial diagnosis from QEMU | |