| pas-ha[m] | hi manila cores, I'd like to bring your attention to https://bugs.launchpad.net/ironic/+bug/2030976/comments/42 | 07:22 |
|---|---|---|
| pas-ha[m] | recent change in oslo.messaging is breaking Manila due to a certain bug in manila's subclass of RequestContext (that you guys inherited from cinder which is also affected) | 07:24 |
| pas-ha[m] | effectively for manila it means it can not send any oslo.messaging notifications, which e.g. leads to HTTP 500 on share type create (the easiest to test if you install oslo.messaging from git) | 07:25 |
| pas-ha[m] | fixes are here https://review.opendev.org/q/Ifb660688c590fb1c32cbe0ae03e0ea444f19cbb8 | 07:25 |
| pas-ha[m] | and will have to be backported to where the oslo.messaging patch https://review.opendev.org/q/Ic94323658c89df1c1ff32f511ca23502317d0f00 was or is about to be backported | 07:26 |
| opendevreview | kiran pawar proposed openstack/manila master: Fix count in the response of shares/snapshots list API https://review.opendev.org/c/openstack/manila/+/893345 | 10:00 |
| opendevreview | Carlos Eduardo proposed openstack/python-manilaclient master: Bump tox version https://review.opendev.org/c/openstack/python-manilaclient/+/893234 | 10:23 |
| carloss | o/ hi pas-ha[m] thanks for the fix | 10:32 |
| carloss | I'll take a look | 10:32 |
| opendevreview | Merged openstack/python-manilaclient master: Bump tox version https://review.opendev.org/c/openstack/python-manilaclient/+/893234 | 12:12 |
| opendevreview | Thiago José de Andrade Alvoravel proposed openstack/manila-tempest-plugin master: Add manila.conf parameters for NetApp Active IQ weigher https://review.opendev.org/c/openstack/manila-tempest-plugin/+/892010 | 13:01 |
| opendevreview | Merged openstack/python-manilaclient master: Implement share backup https://review.opendev.org/c/openstack/python-manilaclient/+/344671 | 14:30 |
| opendevreview | Jay Faulkner proposed openstack/manila stable/2023.1: Add default to read_deleted in context's from_dict https://review.opendev.org/c/openstack/manila/+/893256 | 14:31 |
| opendevreview | Jay Faulkner proposed openstack/manila stable/zed: Add default to read_deleted in context's from_dict https://review.opendev.org/c/openstack/manila/+/893257 | 14:32 |
| opendevreview | Jay Faulkner proposed openstack/manila stable/yoga: Add default to read_deleted in context's from_dict https://review.opendev.org/c/openstack/manila/+/893260 | 14:32 |
| opendevreview | Jay Faulkner proposed openstack/manila stable/xena: Add default to read_deleted in context's from_dict https://review.opendev.org/c/openstack/manila/+/893381 | 14:32 |
| *** dviroel_ is now known as dviroel | 14:59 | |
| felipe_rodrigues | Hi guys... NetApp CI is failing, it seems that this change that was recently merged broke it (guess) | 16:43 |
| felipe_rodrigues | https://review.opendev.org/c/openstack/manila/+/883655 | 16:43 |
| felipe_rodrigues | If you see, NetApp CI didn't pass on the merged patch.. The scheduler failed excatly with the added lines: | 16:44 |
| felipe_rodrigues | Aug 16 09:48:23.769685 ubuntu manila-scheduler[106143]: ERROR oslo_messaging.rpc.server File "/opt/stack/manila/manila/db/sqlalchemy/api.py", line 1926, in share_instance_sizes_sum_by_host | 16:44 |
| felipe_rodrigues | Aug 16 09:48:23.769685 ubuntu manila-scheduler[106143]: ERROR oslo_messaging.rpc.server AttributeError: type object 'ShareInstance' has no attribute 'share' | 16:45 |
| felipe_rodrigues | The patch added this line with joining the "share" field of the ShareInstance model: | 16:46 |
| felipe_rodrigues | https://review.opendev.org/c/openstack/manila/+/883655/3/manila/db/sqlalchemy/api.py#1912 | 16:46 |
| felipe_rodrigues | From my view, "ShareInstance" model hasn't any "share" field.. is it right ? | 16:46 |
| felipe_rodrigues | The community CI passed but my guess is that it is not testing the merged line, the code is only reached if: "self.thin_provisioning and self.provisioned_capacity_gb is None" | 16:49 |
| felipe_rodrigues | NetApp code has that condition, it is now broken. | 16:49 |
| felipe_rodrigues | carloss and gouthamr and the patch owner, can you help us ? Am I missing something ? Since we are close to the merge date, I think it seems urgent | 16:51 |
| carloss | felipe_rodrigues: hi I think you're reading it correctly - shareinstance table only has a share_id | 16:54 |
| carloss | maybe we should do that join differently | 16:55 |
| carloss | carthaca: was this change working properly on your CI/env? | 16:55 |
| felipe_rodrigues | Can we revert this merge ? So we can unblock netapp CI and we can work on this scheduler improvement with more info and more detail.. | 16:56 |
| carloss | https://www.irccloud.com/pastebin/1hPxjaPq/ | 16:56 |
| carloss | I believe fixing would be the right thing rather than reverting IMHO | 16:57 |
| felipe_rodrigues | I guess the idea should be as you pointed.. however, we have to test manually, wait for CI and get the approvals from patch owner and so on.. it could take long.. | 16:57 |
| felipe_rodrigues | My idea is more like: rever the patch, let to the patch's owner the fix and we can merge it when it has the correct solution.. with at least a QA and NetApp CI happy.. | 16:59 |
| felipe_rodrigues | Revert the patch is faster.. fixiing the problem consumes time.. tomorrow would be the FF.. My ActiveIQ patch is not passing because of it.. So, I would say that we could speed up the solution by reverting | 17:00 |
| felipe_rodrigues | just an idea | 17:00 |
| felipe_rodrigues | s/rever/severt/ | 17:01 |
| felipe_rodrigues | s/rever/revert/ | 17:01 |
| gmann | carloss: reminder https://review.opendev.org/c/openstack/releases/+/885707 | 17:01 |
| carloss | felipe_rodrigues: we wouldn't need to get the owner's approval in case the NetApp CI is able to QA it and ensure it works | 17:05 |
| carloss | anyway, both approaches work for me | 17:05 |
| carloss | is the netapp ci healthy? wondering because I didn't see its vote on the change, and we couldn't see the failure. | 17:06 |
| carloss | o/ gmann on it, sorry for the wait | 17:06 |
| gmann | carloss: np! thanks | 17:07 |
| caiquemello[m] | Hi carlooss, actually the netapp ci voted in the change. | 17:10 |
| felipe_rodrigues | up to u carloss.. I am ok with both.. my only concern is about merging the ActiveIQ weigher on time.. I mean, it is blocked while this problem is part of the master.. | 17:11 |
| caiquemello[m] | both dhss-no and dhss failed for the same reason pointed by felipe | 17:12 |
| carloss | caiquemello[m]: it isn't showing up in the zuul summary tab, and if it isn't, some comments are swallowed by gerrit sometimes, unless you expand them | 17:12 |
| carloss | wich was the case. | 17:12 |
| carloss | which* | 17:12 |
| felipe_rodrigues | no worries carloss.. which approach do you prefer ? | 17:12 |
| caiquemello[m] | carloss: ack | 17:14 |
| carloss | gmann: might be a "stupid" question, but I was checking the hash on the commit (9cefa379a5b482f19e4f821986b1d92ec5ac84c5) and it doesn't match with a release thing, it matches with: https://review.opendev.org/c/openstack/cyborg-tempest-plugin/+/823330 instead... so, is the hash correct? | 17:20 |
| gmann | carloss: oh checking | 17:21 |
| carloss | felipe_rodrigues: considering time constraints, let's revert the change. the author has until rc1 to fix the issues pointed out | 17:21 |
| carloss | please comment the issue you found in the change and also launchpad | 17:21 |
| felipe_rodrigues | nice | 17:21 |
| felipe_rodrigues | thks | 17:21 |
| opendevreview | Carlos Eduardo proposed openstack/manila master: Revert "Improve scheduler performance when thin provisioning" https://review.opendev.org/c/openstack/manila/+/893384 | 17:23 |
| gmann | carloss: the hash is this one https://github.com/openstack/manila-tempest-plugin/commit/9cefa379a5b482f19e4f821986b1d92ec5ac84c5 | 17:23 |
| carloss | ah, thanks gmann - sorry, looked only on gerrit, I forgot to look at the github repo for the hashes 🤦♂️ | 17:29 |
| carloss | lgtm, sorry for the wait | 17:29 |
| gmann | carloss: np!, you need to take rest :). thanks btw | 17:30 |
| carloss | haha | 17:30 |
| carloss | np :) | 17:30 |
| opendevreview | Merged openstack/python-manilaclient master: Deprecate "manila" CLI client https://review.opendev.org/c/openstack/python-manilaclient/+/879130 | 18:22 |
| opendevreview | Brian Rosmaita proposed openstack/manila master: Fix openstack-tox-py311 job https://review.opendev.org/c/openstack/manila/+/893427 | 18:44 |
| opendevreview | Merged openstack/manila master: Add default to read_deleted in context's from_dict https://review.opendev.org/c/openstack/manila/+/893223 | 19:37 |
| opendevreview | Thiago José de Andrade Alvoravel proposed openstack/manila-tempest-plugin master: Add manila.conf parameters for NetApp Active IQ weigher https://review.opendev.org/c/openstack/manila-tempest-plugin/+/892010 | 19:58 |
| opendevreview | Merged openstack/manila master: Revert "Improve scheduler performance when thin provisioning" https://review.opendev.org/c/openstack/manila/+/893384 | 20:14 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!