| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2021-12-10 | |||
| 11:42:30 | sean-k-mooney | that might be too early | |
| 11:42:39 | sean-k-mooney | lookign at the logs its installed as part of installing ceph | |
| 11:42:47 | sean-k-mooney | so i might need to fix this in the ceph plugin instead | |
| 11:42:59 | sean-k-mooney | its pulled in by ++ functions-common:apt_get:1207 : sudo DEBIAN_FRONTEND=noninteractive http_proxy= https_proxy= no_proxy= apt-get --option Dpkg::Options::=--force-confold --assume-yes install ceph libnss3-tools python3-rados python3-rbd | |
| 11:43:11 | gibi | that can explain why we only see this breaking the ceph job only | |
| 11:43:22 | sean-k-mooney | ill quickly add the same to the ceph plugin | |
| 11:43:42 | gibi | this is what I call teamwork :) | |
| 11:49:41 | sean-k-mooney | im going to push a patch that may not work as it might unistalll ceph again but if that is what happens we can just remove the egg infor files to work aroudn it as we do in teh fixup fucntion | |
| 11:50:23 | gibi | OK | |
| 11:50:51 | sean-k-mooney | https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/821396 | |
| 11:52:28 | sean-k-mooney | im hoping its just a recommended package rather then a required one for those packages | |
| 11:52:57 | sean-k-mooney | actully no it looks like its required | |
| 11:53:11 | sean-k-mooney | it snot in the suggested or recommeded section so ill have to do it the other way | |
| 12:01:57 | sean-k-mooney | ok ill try and check back on that leater but hopfully that will fix it | |
| 12:03:29 | sean-k-mooney | gibi: i pushed the first draft of the healthcheck spec yesterday evenign by the way https://review.opendev.org/c/openstack/nova-specs/+/821279 if you are interested. it defnitly need to be reworked since half of it i wrote 3 weeks ago and i just finished off the other section last night but most of the contrent is at least there | |
| 12:03:49 | gibi | sean-k-mooney: ack, I will read it today | |
| 12:04:48 | sean-k-mooney | cool im going to start POCing the code next week and ill use that to also help inform the design | |
| 12:54:47 | opendevreview | mitya-eremeev-2 proposed openstack/nova master: Delete bogus attachments. https://review.opendev.org/c/openstack/nova/+/820935 | |
| 13:35:30 | opendevreview | Pierre-Samuel Le Stang proposed openstack/nova-specs master: Implements: blueprint soft-delete-instance-actions https://review.opendev.org/c/openstack/nova-specs/+/821387 | |
| 13:40:41 | AlexShishebarov | Hello. I wanted to ask a question about the bug I reported. | |
| 13:41:19 | AlexShishebarov | https://bugs.launchpad.net/nova/+bug/1951623 This bug is that during live migration of a VM with a disabled port (to a neutron), the migration stopes with the error Timed out waiting for events “network-vif-plugged” (for disabled port). Will the change https://review.opendev.org/c/openstack/nova/+/767368 fix this problem? | |
| 14:06:06 | sean-k-mooney | AlexShishebarov: this is currently being fixed | |
| 14:06:27 | sean-k-mooney | AlexShishebarov: https://review.opendev.org/c/openstack/nova/+/819494 shoudl correct the issue | |
| 14:08:28 | sean-k-mooney | https://review.opendev.org/c/openstack/nova/+/767368 will not | |
| 14:38:31 | AlexShishebarov | sean-k-mooney: According to the code does not resolve this bug (in my opinion). Nova will continue to wait for events from the neutron. But events from a neutron will not come for disabled port. https://review.opendev.org/c/openstack/nova/+/819494 nova/network/model.py(line 556) method get_live_migration_plug_time_events | |
| 14:43:50 | sean-k-mooney | we shoudl be waiting for events for live migration | |
| 14:43:58 | sean-k-mooney | for cold migration the even moved | |
| 14:44:15 | sean-k-mooney | it went form bind time to plug tiem on revert | |
| 14:45:49 | sean-k-mooney | you are correct that we shoudl skip disabeld ports | |
| 14:46:02 | sean-k-mooney | well kind of | |
| 14:46:23 | sean-k-mooney | it shoudl not actully be required as neuton shoudl be sending the event but they dont currntly | |
| 14:46:55 | sean-k-mooney | we can skip them to work around the neutron bug but neutron shold be send network-vif-plugged event even for disabled ports | |
| 14:48:07 | sean-k-mooney | for now we shoudl be checking the vif status here https://review.opendev.org/c/openstack/nova/+/819494/5/nova/network/model.py#494 | |
| 14:53:20 | AlexShishebarov | I tried to solve the problem this way, but the nova saves the port state as "active" even if it is turned off to a neutron. https://github.com/openstack/nova/blob/master/nova/network/neutron.py#L3183-L3185 | |
| 14:55:39 | AlexShishebarov | in neutron | |
| 15:33:27 | opendevreview | Balazs Gibizer proposed openstack/nova master: functional: Add reproducer for #1907775 https://review.opendev.org/c/openstack/nova/+/766771 | |
| 15:33:43 | gibi | stephenfin: I hope you are OK that I refreshed ^^ | |
| 15:33:59 | gibi | I would like to add the rejection to the aggregate API to fix the bug | |
| 15:34:07 | gibi | but I figured out that we have more cases to handle | |
| 15:34:11 | gibi | so I added extra tests | |
| 15:51:51 | opendevreview | sean mooney proposed openstack/nova-specs master: Add spec for external cpu power managment https://review.opendev.org/c/openstack/nova-specs/+/821419 | |
| 17:06:31 | sean-k-mooney | gibi: it looks like https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/821396/ works | |
| 17:06:44 | gibi | sean-k-mooney: awesome | |
| 17:07:10 | sean-k-mooney | there are some test failerues in the non voting jobs but they are unrelated | |
| 17:07:23 | sean-k-mooney | who normally reivews that repo? do you know? | |
| 17:08:14 | sean-k-mooney | i might check on the qa channel | |
| 17:10:36 | gibi | I don't. lyarwood might know but he is already on PTO | |
| 17:13:52 | sean-k-mooney | ya i just left a message on #openstack-qa | |
| 17:14:07 | sean-k-mooney | honestly it proably wont be seen until moday at this point | |
| 17:14:14 | gibi | gmann: ^^ | |
| 17:14:28 | gibi | gmann: I think you have +2 rights in devstack-plugin-ceph | |
| 17:14:29 | opendevreview | Balazs Gibizer proposed openstack/nova master: functional: Add reproducer for #1907775 https://review.opendev.org/c/openstack/nova/+/766771 | |
| 17:14:30 | opendevreview | Balazs Gibizer proposed openstack/nova master: Reject AZ changes during aggregate add / remove host https://review.opendev.org/c/openstack/nova/+/821423 | |
| 17:17:21 | sean-k-mooney | ya anyone in devstack-core or qa-release or cinder-core | |
| 17:17:23 | sean-k-mooney | https://review.opendev.org/admin/groups/8e51d8f39e33d932a40f03060b861db57a3d25d4,members | |
| 17:19:30 | sean-k-mooney | on one hand its a gate blocker on the ohter hand its friday evening and its not near a milestone so it can proably wait till monday without it being a major problem | |
| 17:20:09 | sean-k-mooney | i can send a mail to the list just to keep everyone in the loop | |
| 17:21:21 | opendevreview | Artom Lifshitz proposed openstack/nova master: DNM: Debug whitebox RBD direct download test failures https://review.opendev.org/c/openstack/nova/+/821424 | |
| 17:22:15 | gmann | sean-k-mooney: gibi checking | |
| 17:22:21 | gibi | thanks! | |
| 17:32:53 | gibi | sean-k-mooney: left initial feedback on the healtcheck spec | |
| 17:32:59 | gibi | and that I will end my week | |
| 17:33:07 | gibi | see you next week o/ | |
| 17:35:32 | gmann | gibi: sean-k-mooney did not get why it started failing now? is there any recent change triggered this? | |
| 17:43:10 | opendevreview | Ghanshyam proposed openstack/nova master: DNM: testing nova-ceph-multistore fix https://review.opendev.org/c/openstack/nova/+/821435 | |
| 17:47:51 | sean-k-mooney | gmann: not that i know if i suspect either neutron started requrieing a new version of python3-logging or there was a cpeh package update upstream | |
| 17:54:26 | gmann | i can see only in l-c requirements which should not be installed by devstack | |
| 17:56:46 | sean-k-mooney | ya i was checking on codeasearch | |
| 17:56:52 | sean-k-mooney | not sure what is pulling it in to be honest | |
| 17:59:30 | sean-k-mooney | sean@cloud:~$ apt-cache rdepends python3-logutils | |
| 17:59:32 | sean-k-mooney | python3-logutils | |
| 17:59:34 | sean-k-mooney | Reverse Depends: | |
| 17:59:36 | sean-k-mooney | python3-pecan | |
| 17:59:38 | sean-k-mooney | changeme | |
| 17:59:47 | sean-k-mooney | so its comiy for either python3-pecan or changeme | |
| 17:59:56 | sean-k-mooney | likely python3-pecan | |
| 18:02:10 | sean-k-mooney | ya so thye had a release 2 days ago | |
| 18:02:12 | sean-k-mooney | https://pypi.org/project/pecan/ | |
| 18:02:17 | sean-k-mooney | and its used by neturon | |
| 18:02:57 | sean-k-mooney | gmann: and we udpated the upper constratints this morning https://github.com/openstack/requirements/commit/e74f91e250d127ff825dad5219527d62efaeb740 | |
| 18:03:00 | gmann | humm, it was there in pecan 1.4.1 which constraints updated in requirement yesterday https://github.com/pecan/pecan/blob/1.4.1/requirements.txt | |
| 18:03:10 | sean-k-mooney | yep | |
| 18:03:16 | sean-k-mooney | so that is likely what broke it | |
| 18:03:22 | gmann | yeah but there is no change in previous version of pecan on logutils | |
| 18:03:35 | sean-k-mooney | https://github.com/pecan/pecan/blob/master/requirements.txt#L6 | |
| 18:03:43 | sean-k-mooney | the require logutile >=3 | |
| 18:03:58 | gmann | yeah that is there in old version too https://github.com/pecan/pecan/blob/1.3.3/requirements.txt | |
| 18:04:07 | sean-k-mooney | ya 8 years | |
| 18:04:47 | sean-k-mooney | im not sure but that is the best i can come up with the version that is shipped in ubuntu is 0.3.4 i think | |
| 18:05:09 | sean-k-mooney | 0.3.3-7 | |
| 18:05:23 | sean-k-mooney | so ghatt should work | |
| 18:05:33 | gmann | I am trying to check the passing nova-ceph-multistore job and it was not pulled there somwhoe https://zuul.openstack.org/build/590c5996ca1b402486bfe1c7e1d08535/log/job-output.txt | |
| 18:07:29 | sean-k-mooney | its likely a ubuntu or cpeh packageing change | |
| 18:07:39 | sean-k-mooney | to we use the disto ceph or ceph form ceph.com | |
| 18:08:32 | sean-k-mooney | oh | |
| 18:08:38 | sean-k-mooney | the ceph pcakages changed | |
| 18:08:40 | sean-k-mooney | CEPH_PACKAGES='ceph libnss3-tools python3-rados python3-rbd' | |
| 18:08:43 | sean-k-mooney | on passin g | |
| 18:10:34 | sean-k-mooney | actully they end up being the same | |
| 18:10:36 | sean-k-mooney | CEPH_PACKAGES='ceph libnss3-tools python3-rados python3-rbd' | |