15:01:13 <gouthamr> #startmeeting manila
15:01:15 <openstack> Meeting started Thu Apr  2 15:01:13 2020 UTC and is due to finish in 60 minutes.  The chair is gouthamr. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:16 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:18 <openstack> The meeting name has been set to 'manila'
15:01:19 <lseki> o/
15:01:20 <dviroel> o/
15:01:22 <danielarthurt> o/
15:01:28 <tbarron> hi
15:01:29 <carthaca> Hi
15:01:31 <andrebeltrami> hello!
15:01:32 <carloss> hi :)
15:01:36 <vkmc> o/
15:01:46 <gouthamr> courtesy ping: xyang toabctl ganso amito lseki
15:01:55 <ganso> o/
15:01:58 <gouthamr> hey lseki
15:02:09 <gouthamr> Agenda: https://wiki.openstack.org/wiki/Manila/Meetings#Next_meeting
15:02:25 <tbarron> ganso! I saw you in cinder and thought you might love them more than us.
15:02:27 <gouthamr> hello everyone o/
15:02:30 * tbarron is relieved
15:02:30 <vhari> o/
15:02:46 <ganso> tbarron: rofl
15:02:49 <lseki> lol
15:02:53 <gouthamr> #topic Announcements
15:03:29 <gouthamr> I hope everyone is staying home and staying healthy
15:03:47 <gouthamr> We're quietly approaching feature freeze for the Ussuri release
15:03:52 <gouthamr> that's exactly in one week
15:03:58 <gouthamr> #link: https://releases.openstack.org/ussuri/schedule.html
15:04:21 * ganso is in a conflicting meeting
15:05:14 <gouthamr> we'll discuss patches in a bit, but, this milestone is a freeze for all new features/api changes - we'll also be making releases for python-manilaclient and manila-ui
15:05:38 <gouthamr> ganso: ack
15:06:30 <gouthamr> i'd like to not have any exceptions - but, we're living in exceptional times, so if you feel like some work of yours is going to be delayed, please let me know either on irc or via the openstack-discuss mailing list
15:07:14 <gouthamr> no other announcements for today, it's busy around here :)
15:07:26 <gouthamr> anyone else got any?
15:08:04 <gouthamr> #topic Tracking our work
15:08:10 <gouthamr> #link https://etherpad.openstack.org/p/manila-ussuri-review-focus
15:08:53 <gouthamr> lets go down that list
15:08:53 <gouthamr> Support query user message by timestamp
15:09:20 <gouthamr> thanks for reviewing this, i see two patches workflowed
15:09:40 <gouthamr> #link: https://review.opendev.org/#/c/716848/
15:09:40 <gouthamr> haixin added a test patch:
15:10:02 <gouthamr> no rush to get this one in, we've time beyond feature freeze to merge tests too
15:10:11 <gouthamr> [Unity] Manage/unmanage share server/share/snap
15:10:31 <tbarron> unity CI seems problematic
15:11:06 <gouthamr> yes, dviroel suggested a fix yesterday, i'm not sure how far that's progressing
15:11:31 <tbarron> or rather, maybe it is working, catching issues since the code under review is changing in response
15:12:07 <gouthamr> ack, that too - its a good thing to insist for tempest test coverage
15:12:16 <tbarron> gouthamr: I recommend that we not rush merge but allow exception if needed *if* consistent effort continues to be applied
15:12:23 <tbarron> it's an isolated feature
15:12:28 <tbarron> back end specific
15:12:38 <gouthamr> yeah, we'll keep this at the top of the backlog
15:12:42 <dviroel> +1
15:12:50 <gouthamr> Graduate share groups feature
15:12:55 <gouthamr> #link https://review.opendev.org/#/q/topic:bp/graduate-share-groups-feature+(status:open+OR+status:merged)
15:13:19 <gouthamr> I haven't personally gotten to this, but, it's been up there for a bit
15:13:42 <gouthamr> oh boy, looks like we've had no eyes besides carloss
15:13:55 <gouthamr> dviroel vkmc - can you please take a look
15:14:16 <vkmc> gouthamr, will do
15:14:22 <dviroel> gouthamr: sure
15:14:26 <gouthamr> awesome thank you
15:14:29 <carloss> gouthamr: vkmc reviewed the manila one, I have addressed the answers
15:14:40 <carloss> ty vkmc dviroel
15:14:44 <vkmc> thanks carloss, I'll review again
15:14:55 <gouthamr> ah i missed that, ty vkmc carloss
15:14:59 <gouthamr> Tenant based quotas for share replicas and replica sizes
15:15:07 <gouthamr> #link https://review.opendev.org/#/q/topic:bp/limit-share-replicas-per-share+(status:open+OR+status:merged)
15:15:39 <gouthamr> this is proceeding well i presume, you'll need to rebase because of the microversion bump
15:15:48 <dviroel> currently reviewing these changes, also did some QA that I'll comment on the etherpad
15:16:02 <carloss> I have got a new review from dviroel few minutes ago. will address it today. thanks for the review and qa dviroel
15:16:12 <gouthamr> good stuff...
15:16:23 <carloss> s/will address it today/will address the answers today
15:16:24 <gouthamr> Create share from snapshot in another pool or backend
15:16:43 <gouthamr> #link https://review.opendev.org/#/q/topic:bp/create-share-from-snapshot-in-another-pool-or-backend+(status:open+OR+status:merged)
15:17:10 <gouthamr> some good conversations happened wrt this one
15:17:52 <dviroel> I've uploaded a new patch on the core change, adding the check that we talked yesterday
15:17:53 <gouthamr> vkmc, tbarron, i know there were plans to build on this for cephfs - so can you please take a look as well?
15:18:07 <tbarron> +1
15:18:59 <gouthamr> cool dviroel - lets aim to merge early next week, if we can satisfy ourselves
15:19:12 <dviroel> gouthamr: ack, ty
15:19:33 <gouthamr> there's a driver implementation from NetApp:
15:19:35 <gouthamr> #link https://review.opendev.org/#/c/712642/
15:19:41 <gouthamr> it's related
15:20:03 <tbarron> yeah, kinda part of the same bunch
15:20:16 <gouthamr> cool, that's all from this pile
15:20:53 <gouthamr> do we have any python-manilaclient changes and manila-ui changes besides the ones called out above, that need to be there in our ussuri release?
15:21:51 <gouthamr> i know we're still looking at a bunch of changes maaritamm proposed, we can opportunistically get those into this release, but, i don't think we need to rush them all through next week
15:22:02 <gouthamr> (o/ maaritamm)
15:22:31 <gouthamr> also gather that the manila-ui gate's currently broken
15:22:47 <tbarron> +1
15:23:30 <gouthamr> huh, tenant quotas?
15:23:57 <gouthamr> oh well, i'll report a launchpad bug and find vkmc
15:24:00 <gouthamr> i mean owners
15:24:19 <gouthamr> :)
15:24:20 <vkmc> haha
15:24:23 <tbarron> we had ui breakage before when core horizon did something with  quotas
15:24:36 <vkmc> I'm checking out that one... I'm digging into config files for the gate and latest changes in dependencies
15:24:49 <tbarron> we were relying on horizon internals allegedly, might have some of that still
15:24:57 <vkmc> believe so yes
15:25:05 <vkmc> weird thing is that errors cannot be reproduced locally
15:25:09 <gouthamr> oh
15:25:12 <vkmc> I'm finding really funny stuff on the way as well
15:25:18 <gouthamr> not even running the same unit tests?
15:25:25 <tbarron> jaja
15:25:33 <gouthamr> gremlins
15:25:38 <vkmc> :D
15:25:40 <vkmc> smurfs
15:25:51 <tbarron> gouthamr: i ran same unit tests and don't hit it, but I was on fedora
15:26:03 <vkmc> I ran those on bionic, same we are using in the gate
15:26:30 <tbarron> vkmc++
15:26:39 <vkmc> still no clue though
15:26:45 <vkmc> but will let you know when I find something
15:27:05 <vkmc> currently looking the diff on horizon :)
15:27:10 <tbarron> maybe horizon folks can tell us if there's any recent change on their side that could explain it
15:27:12 <tbarron> jinx
15:27:14 <gouthamr> concerning, thank you for chasing this down vkmc
15:27:37 <tbarron> client change: https://review.opendev.org/#/c/712543/ looks close to merge
15:28:12 <tbarron> andrebeltrami: I think you could make carthaca happy pretty easilly on that one
15:28:32 <gouthamr> it will need to change, tbarron
15:28:43 <dviroel> ^ this one will need to be update, microversion will be bumped
15:28:59 <andrebeltrami> @tbarron for sure!
15:29:03 <tbarron> ok, picky picky picky
15:29:11 <gouthamr> yep, bear with these microversion churns
15:29:43 <gouthamr> cool, please pop anything else that you need reviewed on the review etherpad
15:29:52 * gouthamr speaks to those reading meeting logs too
15:30:08 <gouthamr> #link: https://etherpad.openstack.org/p/manila-ussuri-review-focus (Review Focus Etherpad for Ussuri Feature Freeze)
15:30:37 <gouthamr> moving on
15:30:41 <gouthamr> #topic Driver interface changing with new snapshot cloning behavior
15:30:55 <gouthamr> okay, snapshot cloning, it's a term now
15:31:10 <gouthamr> #link https://review.opendev.org/#/c/709697/
15:31:47 <gouthamr> we discussed this patch, so this is an informational update
15:32:00 <gouthamr> the create_share_from_snapshot driver interface is changing
15:32:18 <gouthamr> the change will be updating all in-tree drivers
15:32:28 <gouthamr> and break any out-of-tree drivers
15:32:42 <gouthamr> we don't know who/how many of them exist
15:33:09 <gouthamr> we don't often make these kind of changes lightly
15:33:52 <gouthamr> dviroel: i think it's fitting to send a note to openstack-discuss when you get a chance regarding this, what do you think?
15:34:20 <dviroel> gouthamr: agree, its a good idea. Will do
15:35:00 <dviroel> gouthamr: lets wait a little bit to see if will be approved :)
15:35:20 <gouthamr> thank you, in general i think it's a good idea to hand down the source/parent share - we should have always done it
15:36:15 <gouthamr> dviroel: ack, we should do around feature freeze
15:36:37 <gouthamr> anything else about $topic?
15:36:46 <gouthamr> #topic Bugs (vhari)
15:36:51 <gouthamr> #link: https://etherpad.openstack.org/p/manila-bug-triage-pad-new (Bug Triage etherpad)
15:36:56 <gouthamr> o/ vhari
15:36:59 <gouthamr> you're up
15:37:01 <vhari> gouthamr, \o
15:37:11 <vhari> gouthamr, let's look at a few new ones
15:37:20 <vhari> #link https://bugs.launchpad.net/manila/+bug/1870280
15:37:21 <openstack> Launchpad bug 1870280 in Manila "[RFE] add user message for share creation failure because of no share-type" [Undecided,New]
15:38:01 <vkmc> heh
15:38:05 <gouthamr> hmm, its a bug that we don't currently have a user message for this
15:38:15 <vhari> gouthamr, we also hit this  in our setup recently
15:38:55 <gouthamr> its an classic asynchronous failure, and a non-privileged user is left wondering what they did wrong
15:40:11 <gouthamr> you could think the bug's in two places: lack of user message
15:40:27 <gouthamr> and python-manilaclient not preventing you from this issue
15:40:47 <gouthamr> manila-ui sort of prevents you from hitting this
15:41:13 <gouthamr> it reads share types from the server and makes it mandatory to pick one when creating a share
15:42:00 <gouthamr> we could argue that it's probably unhelpful if there are no share types in that drop down, head scratcher - but, at least you don't wonder why your shares aren't going to error
15:42:07 <gouthamr> with no explanation
15:42:52 <gouthamr> vhari: lets triage this - i think its medium priority, and a low-hanging-fruit
15:43:11 <vhari> gouthamr, ack
15:43:16 <ganso> I don't see why we can't fail that synchronously
15:43:30 <gouthamr> only difference is i think its a clear bug, not an RFE on manila
15:44:01 <gouthamr> ganso: we could, but it'd involve slowing down the POST /shares API to retrieve share types
15:44:57 <gouthamr> ganso: we don't commit the default share type to the database, so all share types must be retrieved and filtered
15:45:03 <ganso> gouthamr: yea, the API does a bunch of DB queries today to perform validations. This would be one more
15:45:39 <gouthamr> ganso: true, this was my initial assumption looking at the issue
15:45:55 <ganso> gouthamr: oh I see there is no way the DB can know which share type is the default, so the query is not efficient
15:46:52 <gouthamr> yes, it's possible to add that check - but, in a multi-tenant cloud, this behavior may be intentional
15:47:02 <gouthamr> cloud administrators may not have a default share type
15:47:23 <ganso> well, since changing the default share type currently requires a restart, I don't see why we cannot save it in the database. Even if we change it to not require a restart later we could trigger an update whenever it would detect and apply the config change
15:48:08 <gouthamr> ganso: hmm, as a share type attr? or a separate table?
15:48:09 <ganso> gouthamr: that's even better for that scenario, we would fail immediately if the share type is not specified
15:48:25 <ganso> gouthamr: share type attr sounds fine to me
15:49:20 <gouthamr> ganso: it'd be a multi-row update, and you're assuming that config is changed in all places at the same time?
15:50:00 <gouthamr> we're slipping into design discussion, but i think we're onto something ganso
15:50:12 <gouthamr> lets update the bug with these notes
15:50:31 <gouthamr> i'll hold off on the "low-hanging-fruit" classification until we have arrived at a conclusion
15:51:12 <gouthamr> vhari: perhaps pencil this in for revisiting in a few weeks, based off the discussion on the bug?
15:51:27 <vhari> gouthamr, makes sense ..
15:51:45 <vhari> moving on.. #link https://bugs.launchpad.net/manila/+bug/1869712
15:51:46 <openstack> Launchpad bug 1869712 in Manila "Increased schedule time for non thin_provisioned backends" [Medium,New] - Assigned to Jose Castro Leon (jose-castro-leon)
15:52:23 <vhari> this is new and has a proposed fix
15:52:40 <gouthamr> vhari: this one has been triaged, we had a fix, but had to be reverted
15:52:43 <vhari> floating out in case additional discussions
15:52:43 <gouthamr> #link https://review.opendev.org/#/c/716683/
15:52:57 <vhari> gouthamr, ack
15:53:00 <gouthamr> the revert hasn't merged yet, it's been stuck in the gate
15:53:43 <gouthamr> josecastroleon will be reworking the solution based off a finding regarding some share backend capabilities being reported as either booleans or lists
15:54:50 <vhari> gouthamr, will add note to the bug
15:55:06 <vhari> gouthamr, that's a wrap for bugs - more next week
15:55:37 <gouthamr> thank you vhari
15:55:42 <gouthamr> #topic Open Discussion
15:55:50 <vhari> anytime gouthamr :)
15:56:03 <gouthamr> carthaca: I saw you hit an old bug: https://bugs.launchpad.net/manila/+bug/1475351
15:56:04 <openstack> Launchpad bug 1475351 in Manila "Bug: AttributeError: 'module' object has no attribute 'snapshot_delete'" [Undecided,Confirmed]
15:56:39 <dviroel> was the next one in vhari's list
15:56:52 <gouthamr> timely then
15:57:04 <dviroel> :)
15:57:10 <gouthamr> it's nearly 5 years old - at this point it's a feature
15:57:15 <carthaca> edge case like the bug
15:58:06 <gouthamr> carthaca: ack, i think we can reproduce this bug with faking a db lock timeout exception like the real one you hit
15:58:19 <carthaca> implementing 'delete_snapshot' would be a feature change then
15:58:24 <gouthamr> and it's interesting, it's possible even with share creation
15:58:56 <carthaca> delete_snapshot <> snapshot_delete
15:59:01 <gouthamr> carthaca: haha, nope, i just mean the discussion on that bug kinda dead-ended before you revived it
15:59:35 <carthaca> sure, it is a joke :)
15:59:42 <gouthamr> carthaca: :)
15:59:52 <gouthamr> okay, we'll see if anyone picks up the bug
16:00:02 <gouthamr> we're out of time
16:00:21 <gouthamr> please review stuff for feature freeze
16:00:26 <gouthamr> and stay healthy
16:00:31 <gouthamr> thanks for attending o/
16:00:37 <gouthamr> #endmeeting