| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2022-09-22 | |||
| 16:46:28 | stephenfin | Okay. I suspect re-implementing that as e.g. 'openstack quota set --default --instances $INSTANCES' or 'openstack default quota set --instances $INSTANCES' would make more sense | |
| 16:46:46 | stephenfin | and deprecate (for removal) all references to quota classes | |
| 16:47:11 | stephenfin | Sound reasonable? | |
| 16:47:49 | melwitt | yeah, I think that makes sense. quota classes is how you can change defaults over the API (as opposed to the config options). so maybe people do do that (?) | |
| 16:49:10 | stephenfin | yeah, I've no idea, but at least this would be a little more discoverable/require less historical knowledge | |
| 16:49:26 | stephenfin | one more thing: you can set your own quota on a per project basis. What do we call those quotas? Custom quotas? Overridden quotas? Project-specific quotas? | |
| 16:49:52 | melwitt | sean-k-mooney: I don't think it's the reverse ... i.e if you qemu-img info <volume path> it returns raw if you qemu-img info <snapshot path> it returns qcow2. unless I'm just totally misunderstanding something | |
| 16:50:35 | melwitt | stephenfin: the last one, project quotas | |
| 16:50:43 | stephenfin | ta | |
| 16:52:01 | melwitt | sean-k-mooney: <source file='/opt/stack/data/nova/mnt/896fb15da6036b68a917322e72ebfe57/volume-89113873-5c74-4980-8396-f876b7b5101c'/> vs <source file='/opt/stack/data/nova/mnt/896fb15da6036b68a917322e72ebfe57/volume-89113873-5c74-4980-8396-f876b7b5101c.484f7406-3169-4ea5-afda-a7b4657c4d4f' index='1'/> | |
| 16:53:33 | melwitt | the latter is what the instance points to after the snapshot and that path/file format is qcow2 | |
| 17:38:58 | sean-k-mooney | melwitt: so when we create a shapshot we are then running form the delta disk | |
| 17:39:52 | melwitt | sean-k-mooney: right | |
| 17:42:21 | sean-k-mooney | yes so the volume is not the new file in qcow format | |
| 17:42:28 | sean-k-mooney | and the snapshot is the old file | |
| 17:42:40 | sean-k-mooney | because if i boot a second vm form the snapshot | |
| 17:42:52 | sean-k-mooney | i should really get the old files content | |
| 17:43:07 | sean-k-mooney | and creatign the new voluem shoudl create a second deleta disk | |
| 17:43:12 | melwitt | oh, ok I think I see what you're saying | |
| 17:43:55 | sean-k-mooney | its kind of the reverse of what you woudl expect | |
| 17:44:18 | melwitt | yeah. I have clearly been confused by all of this 😆 | |
| 17:44:29 | sean-k-mooney | normally we upload a new image to glance with the delta form the base file | |
| 17:44:49 | sean-k-mooney | but the base file does not change with glance | |
| 17:44:58 | sean-k-mooney | but with a voluem it writable | |
| 17:45:11 | sean-k-mooney | so the volume becomes the new file | |
| 17:45:18 | sean-k-mooney | and the snapshot is the old file | |
| 17:46:38 | sean-k-mooney | that i think is how we should look at it but maybe that is not how cinder looks at it | |
| 17:47:11 | sean-k-mooney | to me the voluem is the thing attached to the vm and the snapshot is the backing file | |
| 17:48:45 | melwitt | yeah, it is presented that way as in, the instance remains attached to the same volume uuid, even after snapshots | |
| 17:49:00 | melwitt | (when you look at server show, for example) | |
| 17:49:36 | melwitt | that's part of why it confuses me bc it's attached to the volume but then the xml points at the delta | |
| 17:54:01 | sean-k-mooney | yep so that is why we probly need to change the forma on the volume | |
| 17:54:19 | sean-k-mooney | and keep the format of the snapshot at the current romat of the disk | |
| 17:54:41 | sean-k-mooney | a second snapshot will result in the qcow the vm is now using being the snapshot disk | |
| 17:54:52 | sean-k-mooney | so the second snabp shot format will be qcow | |
| 17:54:58 | sean-k-mooney | but that ok | |
| 17:55:10 | sean-k-mooney | the rule is | |
| 17:55:19 | sean-k-mooney | the format of the snapshot is the current volume format | |
| 17:55:33 | sean-k-mooney | and the volmue format after snapstho is always qcow | |
| 17:57:19 | melwitt | sean-k-mooney: ok. so PS1 of my cinder patch was likely the right approach. and the volume extend failure is expected and should be skipped if nfs is being used, until that spec you linked earlier is implemented (?) | |
| 18:00:12 | sean-k-mooney | i belive so | |
| 18:00:33 | sean-k-mooney | the patch you linked me would also proably work as a workaround | |
| 18:00:51 | sean-k-mooney | but i dont belive it would be the right long term solution | |
| 18:00:59 | melwitt | gotcha | |
| 18:11:40 | sean-k-mooney | o/ | |
| 22:23:58 | opendevreview | Merged openstack/placement master: update bindep for ubuntu 22.04 https://review.opendev.org/c/openstack/placement/+/858927 | |
| 22:24:00 | opendevreview | Merged openstack/placement master: Switch to 2023.1 Python3 unit tests and generic template name https://review.opendev.org/c/openstack/placement/+/857901 | |
| 23:07:30 | opendevreview | melanie witt proposed openstack/nova stable/yoga: Unify placement client singleton implementations https://review.opendev.org/c/openstack/nova/+/858997 | |
| 23:07:31 | opendevreview | melanie witt proposed openstack/nova stable/yoga: Avoid n-cond startup abort for keystone failures https://review.opendev.org/c/openstack/nova/+/858998 | |
| 23:12:18 | opendevreview | melanie witt proposed openstack/nova stable/xena: Unify placement client singleton implementations https://review.opendev.org/c/openstack/nova/+/858999 | |
| 23:12:19 | opendevreview | melanie witt proposed openstack/nova stable/xena: Avoid n-cond startup abort for keystone failures https://review.opendev.org/c/openstack/nova/+/859000 | |
| 23:23:41 | opendevreview | melanie witt proposed openstack/nova stable/wallaby: Unify placement client singleton implementations https://review.opendev.org/c/openstack/nova/+/859001 | |
| 23:23:42 | opendevreview | melanie witt proposed openstack/nova stable/wallaby: Avoid n-cond startup abort for keystone failures https://review.opendev.org/c/openstack/nova/+/859002 | |
| #openstack-nova - 2022-09-23 | |||
| 02:46:44 | opendevreview | Brian Rosmaita proposed openstack/nova master: Correct reST markup in config help string https://review.opendev.org/c/openstack/nova/+/859010 | |
| 02:54:59 | opendevreview | Junbo Jiang proposed openstack/nova master: Test overcommit status when choose numa nodes https://review.opendev.org/c/openstack/nova/+/858495 | |
| 03:37:14 | opendevreview | Rajesh Tailor proposed openstack/nova master: Update Availability zone doc page https://review.opendev.org/c/openstack/nova/+/846463 | |
| 03:39:52 | opendevreview | Rajesh Tailor proposed openstack/nova master: Fix typos in nova docs https://review.opendev.org/c/openstack/nova/+/858673 | |
| 03:41:01 | opendevreview | Merged openstack/nova master: Update nova-manage doc page https://review.opendev.org/c/openstack/nova/+/856894 | |
| 10:31:02 | auniyal_ | Hi #openstack-nova | |
| 10:31:14 | auniyal_ | regarding error: service catalog is empty | |
| 10:31:26 | auniyal_ | I am writing a functional test, to create a snapshot of instance whose image has property "os_require_quiesce" and should boot-from-volume. | |
| 10:31:26 | auniyal_ | so while creating snapshot I can test _set_quiesced functionality. | |
| 10:31:26 | auniyal_ | https://opendev.org/openstack/nova/src/branch/master/nova/virt/libvirt/driver.py#L3238 | |
| 10:31:26 | auniyal_ | I have this functional test - https://paste.opendev.org/show/bY6ZjvBzDO4QXbzrpQ70/ | |
| 10:31:26 | auniyal_ | It has 2 WAY's | |
| 10:31:30 | auniyal_ | WAY 1: boot from volume as I want to write, but its failing and giving error: https://paste.opendev.org/show/bMRo6x3aDgKmgDbQbadr/ | |
| 10:31:33 | auniyal_ | WAY 2: use CINDER fixture to get volume, Error: https://paste.opendev.org/show/bj553NIDlVAw7sKkVZfF/ | |
| 10:32:51 | sean-k-mooney | you shoudl be using the cinder fixture yes | |
| 10:33:27 | sean-k-mooney | your not actully usin gthe cidner fixture in this test | |
| 10:33:31 | sean-k-mooney | https://paste.opendev.org/show/bY6ZjvBzDO4QXbzrpQ70/ | |
| 10:33:36 | sean-k-mooney | yoru using a uuid from it | |
| 10:33:53 | sean-k-mooney | but you have not actully used the cinder fixture to provide a fake sinder | |
| 10:33:58 | sean-k-mooney | *cinder | |
| 10:34:06 | auniyal_ | in WAY 2, line 52 | |
| 10:34:07 | sean-k-mooney | right now your test is trying to actully call keystone | |
| 10:34:15 | sean-k-mooney | thats not using the fixture | |
| 10:34:22 | sean-k-mooney | on line 52 | |
| 10:34:32 | sean-k-mooney | its just geting a constnat that is defiend in it | |
| 10:35:33 | auniyal_ | okay, but shouldn't it be treated as volume while instance creation, (fake volume) | |
| 10:36:30 | auniyal_ | or I should add some other property as well | |
| 10:36:56 | sean-k-mooney | add self.useFixture(nova_fixtures.CinderFixture(self)) to the setUp | |
| 10:37:12 | sean-k-mooney | on line 20 | |
| 10:37:36 | sean-k-mooney | and from nova.tests import fixtures as nova_fixtures | |
| 10:38:01 | sean-k-mooney | ah you have that on line 4 | |
| 10:38:32 | sean-k-mooney | to enable a fixture you actullly need to do self.useFixture | |
| 10:39:28 | sean-k-mooney | also why did you comment out base.ServersTestBase, and integrated_helpers.InstanceHelperMixin | |
| 10:40:09 | auniyal_ | actually, CinderFixture is already added in inherited class | |
| 10:40:13 | auniyal_ | in here - https://opendev.org/openstack/nova/src/commit/aad31e6ba489f720f5bdc765c132fd0f059a0329/nova/tests/functional/integrated_helpers.py#L1208 | |
| 10:41:14 | sean-k-mooney | right but your not ment to be inheriting form that | |
| 10:42:02 | auniyal_ | okay, so I should inherit base.ServersTestBase, and integrated_helpers.InstanceHelperMixin | |
| 10:42:08 | auniyal_ | and add CinderFixtute | |
| 10:42:52 | sean-k-mooney | yes as i said a few time in the past integrated_helpers._IntegratedTestBase does not use the fake libvirt implemeattion | |
| 10:43:03 | sean-k-mooney | the bug your working on only happens in the libvirt driver | |
| 10:43:08 | sean-k-mooney | so you cant use that | |
| 10:43:28 | sean-k-mooney | which is why i perviouly told you to use base.ServersTestBase, and integrated_helpers.InstanceHelperMixin | |
| 10:45:08 | auniyal_ | actually, I tried that as well, so thats why, only commented and not removed from test | |
| 10:46:17 | auniyal_ | it failes with - https://paste.opendev.org/show/bL7nGtDSyTzc11Utb0AA/ | |
| 11:02:12 | sean-k-mooney | yes that is a diffent issue | |
| 11:02:33 | sean-k-mooney | we need to mock out the calls to check for secure boot | |
| 11:03:09 | sean-k-mooney | that is at least using the libvirt driver which is good | |