| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2022-07-18 | |||
| 11:38:47 | opendevreview | Merged openstack/os-traits master: Add 'COMPUTE_STORAGE_VIRTIO_FS', 'COMPUTE_MEM_BACKING_FILE' https://review.opendev.org/c/openstack/os-traits/+/832769 | |
| 11:40:30 | sean-k-mooney | Uggla: ill propose a release of os-traits later this week there is one other patch i want to land | |
| 11:41:08 | sean-k-mooney | but that will allow you to use it form pypi once we have done the release and unblock the nova series form merging. | |
| 11:43:09 | sean-k-mooney | well its still WIP but more it will unblock the unit/functional tests form passign in the gate | |
| 11:43:28 | Uggla | sean-k-mooney, cool thx | |
| 11:46:55 | opendevreview | sean mooney proposed openstack/nova master: enable blocked VDPA move operations https://review.opendev.org/c/openstack/nova/+/832330 | |
| 11:51:39 | opendevreview | sean mooney proposed openstack/nova master: enable blocked VDPA move operations https://review.opendev.org/c/openstack/nova/+/832330 | |
| 11:56:57 | opendevreview | Balazs Gibizer proposed openstack/nova master: Gracefully ERROR in _init_instance if vnic_type changed https://review.opendev.org/c/openstack/nova/+/850003 | |
| 11:58:26 | opendevreview | Balazs Gibizer proposed openstack/nova master: Gracefully ERROR in _init_instance if vnic_type changed https://review.opendev.org/c/openstack/nova/+/850003 | |
| 12:01:59 | opendevreview | Balazs Gibizer proposed openstack/nova master: Move uwsgi-gmr reno to the proper place https://review.opendev.org/c/openstack/nova/+/850152 | |
| 12:22:31 | Uggla | gibi, just get a new fresh repo from my code and still having the unit tests failing.... becoming crazy... | |
| 12:23:00 | sean-k-mooney | what version of python are you using and what os | |
| 12:23:43 | Uggla | 3.10.5 | |
| 12:23:52 | Uggla | Fedora 36 | |
| 12:23:59 | sean-k-mooney | i see well that could be the issue | |
| 12:24:06 | sean-k-mooney | 3.10 support is experimental | |
| 12:24:18 | sean-k-mooney | ill see if i get the same issue with 3.10 | |
| 12:24:38 | sean-k-mooney | oh hum | |
| 12:24:43 | sean-k-mooney | i was using 3.10.4 | |
| 12:24:52 | Uggla | but within tox 3.9 is used no ? | |
| 12:25:13 | sean-k-mooney | only if you run tox -e py39 | |
| 12:25:52 | sean-k-mooney | i ran with 3.10.4 so its not 3.10 | |
| 12:26:07 | Uggla | yep this is what I run. | |
| 12:26:44 | sean-k-mooney | well this seams to work on ubuntu so it might be a fedora dep issue | |
| 12:26:56 | sean-k-mooney | although im not sure what would cause that to fail | |
| 12:30:21 | Uggla | sean-k-mooney, it seems it happened to me on Friday after a rebase to split patches... I thought I did something wrong... | |
| 13:07:27 | fungi | sean-k-mooney: did you see my question on https://launchpad.net/bugs/1981813 (just trying to confirm the bug type change was intentional before i add an ossa bugtask)? | |
| 13:10:29 | opendevreview | Manuel Bentele proposed openstack/nova-specs master: Add configuration options to set SPICE compression settings https://review.opendev.org/c/openstack/nova-specs/+/849488 | |
| 13:12:05 | opendevreview | Manuel Bentele proposed openstack/nova master: libvirt: Add configuration options to set SPICE compression settings https://review.opendev.org/c/openstack/nova/+/828675 | |
| 14:28:02 | bauzas | dang. | |
| 14:28:26 | bauzas | https://blueprints.launchpad.net/nova/+spec/configurable-instance-domains from https://specs.openstack.org/openstack/nova-specs/specs/zed/approved/configurable-instance-domains.html is giving me HTTP404 | |
| 14:28:55 | bauzas | looks like artom forgot to create it | |
| 14:29:26 | bauzas | also, I'm a bit surprised we merged a spec within one day for an instance modification, but OK | |
| 14:29:27 | sean-k-mooney | oh well thats an easy fix at least | |
| 14:30:21 | sean-k-mooney | it was basically the same change we did for hostname 2 cyles ago | |
| 14:30:32 | sean-k-mooney | so that is not really contovial at least in my view | |
| 14:30:48 | sean-k-mooney | we proably should have added both orginally | |
| 14:31:38 | bauzas | mmmm | |
| 14:31:57 | bauzas | we'll see whether it would be a problem later | |
| 14:33:29 | sean-k-mooney | if you think there is a design issue we can revert the spec. | |
| 14:33:49 | sean-k-mooney | are you worried we missed something specific | |
| 14:36:38 | bauzas | well | |
| 14:36:52 | bauzas | I'm afraid it's about a large table | |
| 14:37:06 | bauzas | within also a cells v2 DB | |
| 14:37:33 | bauzas | so we need to make sure the upgrade will work for large operators having a lot of instances and cells | |
| 14:37:35 | sean-k-mooney | we had the option to not do a db migration and store it in the instance_system_metadtaa | |
| 14:38:40 | sean-k-mooney | stephenfin: at least seam to prefer treating it like a first class porperty on the instance object and therefor storing it in the instance table | |
| 14:39:20 | sean-k-mooney | so we can revsit that but we should be abel to make additive changes to the db | |
| 14:40:19 | sean-k-mooney | bauzas: stephenfin was suggesting this si a good way to test the new alembic migration flow as its relitivly low risk | |
| 14:40:25 | bauzas | fortunately, we don't need to modify the data | |
| 14:40:38 | sean-k-mooney | right it just a new nullable column | |
| 14:41:31 | bauzas | but yeah, we need to explain the operators that it would be an ALTER for the instances table | |
| 14:41:50 | bauzas | also, making sure we also add it to the shadow table | |
| 14:41:54 | sean-k-mooney | is the release note not sufficent for that | |
| 14:42:00 | bauzas | it will I guess | |
| 14:43:25 | bauzas | sean-k-mooney: I don't see the default policy for seeing the domain attribute value and putting it | |
| 14:43:33 | bauzas | sean-k-mooney: I guess this is enduser ? | |
| 14:45:10 | bauzas | also, we'll change the notification, as you said, so we need to modify the spec to no longer say "none" to notifications impact | |
| 14:45:12 | sean-k-mooney | it woudl admin_or_opwner | |
| 14:45:18 | sean-k-mooney | the same as hostname | |
| 14:46:29 | sean-k-mooney | or project_member in the new RBAC schem for update and project_reader for viewing | |
| 14:46:52 | sean-k-mooney | we can capture this extra detail in a follow up patch | |
| 14:47:06 | sean-k-mooney | can you leave that feedback for atrom on the patch | |
| 14:48:11 | dansmith | oh are we talking about the domain spec? | |
| 14:48:17 | sean-k-mooney | yes | |
| 14:48:49 | dansmith | submitted and merged in less than 24 hours.. that was not cool, IMHO | |
| 14:49:07 | gibi | I take the blame | |
| 14:49:25 | gibi | I was under the impression that it is verry similar to what we already did | |
| 14:49:31 | gibi | about the hostname | |
| 14:49:54 | sean-k-mooney | it is but the question i really have are there outstanding questions that needed to be adressed | |
| 14:49:59 | dansmith | yeah, but that's not a good (enough) excuse :) | |
| 14:50:05 | sean-k-mooney | i did not think there were | |
| 14:50:20 | gibi | dansmith: I see now | |
| 14:50:22 | sean-k-mooney | but if there are as i mentioned above we can revert or adress in a followup | |
| 14:50:39 | dansmith | I've put some general and specific concerns in there | |
| 14:50:40 | sean-k-mooney | do we dissagre directionally on the change | |
| 14:50:53 | dansmith | technically the process is to fast revert when there wasn't time for feedback | |
| 14:51:06 | dansmith | I know there's a deadline at hand, but still... | |
| 14:51:23 | sean-k-mooney | ack well we can follow the spec freeze excpetion process | |
| 14:51:54 | sean-k-mooney | which is ask for exception in team meeting and then basicaly have till end of week or next meeting to agree or punt | |
| 14:54:49 | sean-k-mooney | hum artom does not apper to be on upstream or downstream irc but ill ping them when they are about to readback and review the spec | |
| 14:55:00 | gibi | artom is on PTO this week | |
| 14:55:00 | sean-k-mooney | bauzas: dansmith do either of ye want to propose the revert patch | |
| 14:55:07 | sean-k-mooney | oh | |
| 14:55:09 | sean-k-mooney | ok | |
| 14:55:17 | dansmith | >:( | |
| 14:55:44 | sean-k-mooney | well in that case i can proably recuse myself form the spec review and take it over and adress the feedback | |
| 14:55:52 | sean-k-mooney | if the rest of ye feel ok with reviewing it | |
| 14:56:10 | gibi | sean-k-mooney: works for me | |
| 14:56:16 | sean-k-mooney | or we punt it to next cycle | |
| 14:58:22 | bauzas | I don't know what to say | |
| 14:58:32 | bauzas | the blueprint isn't created yet | |
| 14:58:50 | bauzas | so technically even if the spec is merged, we don't track its progress | |
| 14:59:02 | bauzas | now, about the spec | |
| 14:59:12 | bauzas | I have a few left comments | |
| 14:59:42 | sean-k-mooney | ya sorry i tought we had a check in the ci that the url existied but i guess not i didnt think to check that manually | |
| 14:59:54 | bauzas | my concern being the discrepancy between neutron's dns_domain and nova's domain and how we reconcile them | |
| 15:00:16 | sean-k-mooney | well there isnnt a depency here | |
| 15:00:42 | sean-k-mooney | we optionally can set the port dns_domain if the extension extis for port that are create by nova | |
| 15:00:43 | bauzas | anyway, we have to join some internal meeting | |