| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2021-07-07 | |||
| 13:28:45 | bauzas | ta, looking | |
| 13:28:50 | stephenfin | gmann, lyarwood: either of you care to re-review also. Super trivial ^ | |
| 13:29:18 | opendevreview | Stephen Finucane proposed openstack/nova master: docs: Replace 'nova boot' with 'openstack server create' https://review.opendev.org/c/openstack/nova/+/794007 | |
| 13:29:28 | stephenfin | lyarwood: I also addressed your comments on that doc | |
| 13:48:01 | lyarwood | ack looking | |
| 14:00:33 | lyarwood | stephenfin: I'm sure you still need --bootable when creating the volume | |
| 14:01:12 | stephenfin | Possibly in the past, but I tested it locally with a basic DevStack deployment and you didn't | |
| 14:02:02 | lyarwood | I've just copy pasted from my env, it's failing for me | |
| 14:03:27 | lyarwood | unless cinder is somehow automatically setting this under a later microversion | |
| 14:03:37 | lyarwood | or we are somehow updating the bootable flag | |
| 14:03:43 | lyarwood | I can't see how this ever worked | |
| 14:04:32 | stephenfin | What I was seeing was that the bootable flag would be False in the response returned from the create request, but then True when I did a show call | |
| 14:04:44 | stephenfin | so cinder was changing things under the good | |
| 14:04:47 | stephenfin | *hood | |
| 14:05:15 | stephenfin | I'm pretty sure I traced through cinder to find what was setting it too, but it seems I didn't take notes on the patch like I usually would | |
| 14:05:42 | lyarwood | weird | |
| 14:07:49 | lyarwood | oh wait that's with the --bootable flag | |
| 14:08:02 | lyarwood | haha nice, I bet the osc command does it in two calls or something weird | |
| 14:08:40 | stephenfin | wdym? | |
| 14:10:12 | lyarwood | https://docs.openstack.org/api-ref/block-storage/v3/index.html?expanded=create-a-volume-detail#volumes-volumes - because c-api doesn't take bootable as an arg when creating the volume | |
| 14:12:34 | lyarwood | https://github.com/openstack/python-openstackclient/blob/4f358367986525625157bee164c694e65573c30c/openstackclient/volume/v2/volume.py#L231-L232 and yeah osc creates and then optionally marks it as bootable if you pass the flag | |
| 14:13:48 | lyarwood | https://github.com/openstack/nova/blob/dfda0c0482acf579452742860cec2bdd64ac616a/nova/block_device.py#L685-L686 is where n-api rejects it btw | |
| 14:21:22 | stephenfin | lyarwood: I'm still seeing the same behavior http://paste.openstack.org/show/807231/ | |
| 14:22:12 | stephenfin | So if I create a volume from an image then the initial response from cinder is bootable='false', but this changes to bootable='true' once the volume is fully created | |
| 14:22:30 | lyarwood | yeah osc must be doing that | |
| 14:22:38 | lyarwood | that makes sense if you've passed --image | |
| 14:23:38 | stephenfin | Not that I can see | |
| 14:23:46 | stephenfin | and python-cinderclient isn't doing anything too funky either | |
| 14:23:53 | lyarwood | yeah weird, maybe c-api then, TIL | |
| 14:24:06 | lyarwood | and I wasn't passing --image so that's why I didn't see it | |
| 14:24:24 | stephenfin | ah, yes, I explicitly called that fact out https://review.opendev.org/c/openstack/nova/+/794007/4/doc/source/user/launch-instance-from-volume.rst#265 | |
| 14:24:29 | stephenfin | Cinder makes a volume bootable when ``--image`` parameter is passed. | |
| 14:26:41 | lyarwood | Yup apologies | |
| 14:28:52 | opendevreview | Jay Faulkner proposed openstack/nova stable/wallaby: [ironic] Minimize window for a resource provider to be lost https://review.opendev.org/c/openstack/nova/+/799772 | |
| 14:34:53 | gmann | stephenfin: done, thanks for adding test. | |
| 14:36:36 | stephenfin | ta | |
| 15:22:33 | opendevreview | Lee Yarwood proposed openstack/nova-specs master: Add connection_info to responses of the os-volume_attachments API for admins https://review.opendev.org/c/openstack/nova-specs/+/799850 | |
| 15:44:29 | opendevreview | Lee Yarwood proposed openstack/nova-specs master: Add connection_info to responses of the os-volume_attachments API for admins https://review.opendev.org/c/openstack/nova-specs/+/799850 | |
| 15:48:40 | opendevreview | Balazs Gibizer proposed openstack/nova master: [func test] neutron fixture for extended resource request https://review.opendev.org/c/openstack/nova/+/794306 | |
| 15:48:41 | opendevreview | Balazs Gibizer proposed openstack/nova master: Detect extended_resource_request neutron API extension https://review.opendev.org/c/openstack/nova/+/793618 | |
| 15:50:01 | opendevreview | Balazs Gibizer proposed openstack/nova master: Reject server create with extended resource req https://review.opendev.org/c/openstack/nova/+/793619 | |
| 15:50:02 | opendevreview | Balazs Gibizer proposed openstack/nova master: Reject server operations with extended resource req https://review.opendev.org/c/openstack/nova/+/793620 | |
| 15:52:35 | opendevreview | Balazs Gibizer proposed openstack/nova master: Add same_subtree field to RequestLevelParams https://review.opendev.org/c/openstack/nova/+/791503 | |
| 15:55:23 | opendevreview | Balazs Gibizer proposed openstack/nova master: Bump min placement microversion to 1.36 https://review.opendev.org/c/openstack/nova/+/791504 | |
| 15:57:00 | opendevreview | Balazs Gibizer proposed openstack/nova master: Support same_subtree in allocation_canadidate query https://review.opendev.org/c/openstack/nova/+/791505 | |
| 15:57:02 | opendevreview | Balazs Gibizer proposed openstack/nova master: [func test] refactor assertPortMatchesAllocation https://review.opendev.org/c/openstack/nova/+/792458 | |
| 15:59:40 | opendevreview | Balazs Gibizer proposed openstack/nova master: [func test] refactor asserts in qos tests https://review.opendev.org/c/openstack/nova/+/798930 | |
| 16:01:01 | opendevreview | Balazs Gibizer proposed openstack/nova master: Support the new port resource_request format https://review.opendev.org/c/openstack/nova/+/787208 | |
| 16:02:25 | opendevreview | Balazs Gibizer proposed openstack/nova master: Transfer RequestLevelParams from ports to scheduling https://review.opendev.org/c/openstack/nova/+/791506 | |
| 16:03:47 | opendevreview | Balazs Gibizer proposed openstack/nova master: [func test] ports with both bw and pps resources https://review.opendev.org/c/openstack/nova/+/792394 | |
| 16:05:15 | opendevreview | Balazs Gibizer proposed openstack/nova master: [func test] move unshelve test to the proper place https://review.opendev.org/c/openstack/nova/+/793621 | |
| 16:15:10 | opendevreview | Lee Yarwood proposed openstack/nova-specs master: Add nova-manage commands to show and refresh connection_info https://review.opendev.org/c/openstack/nova-specs/+/799624 | |
| 16:45:42 | stephenfin | zzzeek: If you've any spare time whatsoever this week, I'd appreciate your attention on an issue I'm having during the sqla-migrate -> alembic switch. Basically, my migration code to create shadow tables doesn't seem to be working with in-memory SQLite databases, but I can't figure out how to fix it | |
| 16:45:58 | stephenfin | zzzeek: I have a pretty minimal reproducer here https://github.com/stephenfin/alembic-sqlite-memory-issue | |
| 16:46:28 | stephenfin | zzzeek: and the series to do the conversion for nova ends here, fwiw https://review.opendev.org/c/openstack/nova/+/799530/ | |
| 16:47:16 | stephenfin | zzzeek: I _think_ I'm seeing an issue like the one in the cookbook, but I don't understand the issue well enough yet to say /o\ Any input would be much appreciated https://alembic.sqlalchemy.org/en/latest/cookbook.html#sharing-a-connection-with-a-series-of-migration-commands-and-environments | |
| 17:06:55 | opendevreview | Lee Yarwood proposed openstack/nova-specs master: Store and allow libvirt instance device buses and models to be updated https://review.opendev.org/c/openstack/nova-specs/+/799811 | |
| 17:07:54 | lyarwood | stephenfin: hope you're happy, I pretty much always forget to manually run pep8 in projects outside of Nova now thanks to pre-commit ;) | |
| 17:20:09 | sean-k-mooney | hehe same | |
| 17:27:35 | opendevreview | Merged openstack/nova master: api: Handle invalid volume UUIDs during spawn https://review.opendev.org/c/openstack/nova/+/794006 | |
| 18:26:55 | zzzeek | stephenfin: the issue is a combination of using NullPool with create_engine() combined with the usage of a different connection here: https://github.com/stephenfin/alembic-sqlite-memory-issue/blob/main/foo/migrations/versions/8f2f1571d55b_initial_version.py#L98 | |
| 18:27:18 | zzzeek | stephenfin: when you're in a migration, there's a method op.get_bind(). that gives you the current Connection. use that everywhere for SQL, table reflection, etc | |
| 18:34:07 | zzzeek | stephenfin: I marked it up quite a bit to get you started :) | |
| 19:28:56 | rm_work | anyone seen issues with this error on starting up nova-compute (I'm on wallaby): `greenlet.error: cannot switch to a different thread` | |
| 19:29:01 | rm_work | full traceback: http://paste.openstack.org/show/807240/ | |
| 19:29:19 | sean-k-mooney | is that in the api | |
| 19:29:22 | rm_work | log just shows it trying to do heartbeats and then ... this, and the process locks up | |
| 19:29:28 | rm_work | nova-compute service | |
| 19:29:40 | sean-k-mooney | that should not happen in the compute service | |
| 19:30:02 | sean-k-mooney | we do use real threas but only to interact with a select number of things liek libvirt | |
| 19:30:18 | rm_work | hmm, no VMs even in this system yet, it's essentially a clean install | |
| 19:30:34 | rm_work | just trying to start it up for the first time | |
| 19:30:54 | sean-k-mooney | so i woudl assume this is somehow related to the perodic taks | |
| 19:31:00 | sean-k-mooney | *tasks | |
| 19:31:43 | sean-k-mooney | do you have anything custom set in you nova.conf | |
| 19:31:48 | rm_work | hmmm... what periodics does nova-compute run? can I disable them to check? | |
| 19:31:57 | sean-k-mooney | e.g. are you trying to use any atypical features | |
| 19:32:03 | sean-k-mooney | it runs quite a few | |
| 19:32:19 | sean-k-mooney | i think all or most of them have config options | |
| 19:32:23 | rm_work | It's hard for me to answer that question -- yes, the conf is probably custom, but I don't know enough about nova's features to know which are atypical | |
| 19:32:29 | rm_work | I've tried to trim it down as much as possible | |
| 19:33:12 | sean-k-mooney | i was wondering if you were tweakign any oslo concurancy setting related to threading for example | |
| 19:33:22 | sean-k-mooney | you coudl try setting https://docs.openstack.org/nova/latest/configuration/config.html#DEFAULT.run_external_periodic_tasks to false | |
| 19:33:27 | rm_work | ah this is a nova-compute configured for ironic | |
| 19:33:42 | rm_work | ooo, will try that for testing | |
| 19:34:29 | rm_work | no config in oslo_concurrency besides setting the lock_path | |
| 19:34:55 | sean-k-mooney | oh its ironic hum | |
| 19:35:03 | sean-k-mooney | i dont think that shoudl really change things | |
| 19:35:06 | sean-k-mooney | https://docs.openstack.org/nova/latest/configuration/config.html#DEFAULT.sync_power_state_interval | |
| 19:35:24 | sean-k-mooney | there are several "*_interval" | |
| 19:35:34 | sean-k-mooney | config option which contol the perodic tasks | |
| 19:35:40 | rm_work | hmm, added that `run_external_periodic_tasks` to false, didn't help | |
| 19:35:44 | sean-k-mooney | do you have a more complete errror message | |
| 19:35:54 | sean-k-mooney | so we can see what it was doing right before the error | |
| 19:35:58 | rm_work | yeah secx | |
| 19:38:56 | rm_work | http://paste.openstack.org/show/807242/ | |
| 19:39:26 | sean-k-mooney | oh | |
| 19:39:34 | sean-k-mooney | its right after the heartbeat | |
| 19:39:37 | rm_work | yeah | |
| 19:39:39 | sean-k-mooney | one sec | |