15:01:09 <gouthamr> #startmeeting manila
15:01:10 <openstack> Meeting started Thu Jun 25 15:01:09 2020 UTC and is due to finish in 60 minutes.  The chair is gouthamr. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:11 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
15:01:14 <openstack> The meeting name has been set to 'manila'
15:01:23 <tbarron> hi
15:01:29 <maaritamm> o/
15:01:29 <dviroel> o/
15:01:32 <tbarron> gouthamr never sleeps
15:01:33 <danielarthurt> o/
15:01:36 <carloss> o/
15:01:43 <gouthamr> courtesy ping: ganso vkmc amito lseki andrebeltrami
15:01:48 <vkmc> o/
15:01:52 <gouthamr> haha, got distracted there for a minute
15:02:04 <andrebeltrami> o/
15:02:43 <gouthamr> hello everyone, thanks for joining - looks like we have a pretty light agenda today!
15:02:47 <gouthamr> https://wiki.openstack.org/wiki/Manila/Meetings#Next_meeting
15:02:57 <gouthamr> #topic Announcements
15:03:36 <gouthamr> We're in R-16 on the victoria release schedule
15:03:38 <gouthamr> #link https://releases.openstack.org/victoria/schedule.html (victoria release schedule)
15:03:49 <gouthamr> that's two weeks away from the specs deadline
15:04:11 <gouthamr> ideally we want all specs that we're working on to merge by then
15:04:29 <lseki> o/
15:04:53 <gouthamr> if you anticipate any trouble, please bring it up this week or the next
15:05:51 <gouthamr> that said, reviewers, we've a couple of specs available for review
15:05:54 <gouthamr> https://review.opendev.org/#/q/project:openstack/manila-specs+status:open
15:06:26 <gouthamr> it's time to give these some review attention!
15:07:21 <gouthamr> that's all i had in terms of announcements
15:07:41 <gouthamr> does anyone else have any?
15:08:50 <gouthamr> This change merged over the last week
15:08:59 <gouthamr> #link https://review.opendev.org/#/c/735667/ (Assign TC Liasons for the Victoria Cycle)
15:09:43 <gouthamr> per that, i'd like to welcome gmann, cloudnull as the new liaisons for our project/s
15:10:29 <gouthamr> you know who to contact on the #openstack-tc channel, if tbarron is on a holiday :)
15:10:30 <tbarron> welcome gmann cloudnull !
15:10:33 <dviroel> awesome, welcome!
15:10:44 <carloss> welcome gmann, cloudnull :)
15:11:14 <gouthamr> #topic CI/Gate Status
15:11:30 <gouthamr> just wanted to keep you all in the loop of what the current status of the gate is
15:12:20 <gouthamr> as far as i am aware, there's now no problem with the main branch in manila as well as the stable branches upto stein
15:12:47 <vkmc> gmann++ cloudnull++
15:13:10 <gouthamr> devstack changes for stable/rocky and stable/queens have not merged yet
15:13:24 <gouthamr> you can see the latest status from elod here:
15:13:35 <gouthamr> #link http://lists.openstack.org/pipermail/openstack-discuss/2020-June/015644.html ([all][tc][stable][qa] Grenade testing for Extended Maintenance stable)
15:14:25 <gouthamr> we're discovering that there's at least a few more changes necessary to get things working; so please stay tuned if you have any changes in these older branches
15:14:50 <gouthamr> we've had a degraded status on the python-manilaclient functional job
15:16:02 <gouthamr> a couple of fixes are possible there, i'm investigating the most favorable of them: coverting that job from its legacy definition to native zuulv3
15:16:35 <gouthamr> however that fix isn't the most expedient, there's some trial and error involved
15:17:35 <gouthamr> we'll see if i can get it worked out soon...
15:17:50 <gouthamr> anybody aware of any other issues?
15:18:43 <gouthamr> feel like third party CI was broken a lot too, when reporting on the main branch - i hope its fairly easy to fix these breakages
15:19:26 <gouthamr> if not, please let us know or mail the openstack-discuss list and seek help..
15:20:03 <dviroel> gouthamr: we'll check that later on our ci too
15:20:11 <gouthamr> ty dviroel
15:20:25 <gouthamr> #topic Reviews needing attention
15:21:02 <gouthamr> we're all in the middle of a bunch of things, so we'll skip the etherpad for the week
15:21:45 <gouthamr> but heads up, i'll go through these next week, and see if we have any progress - please update it on the etherpad if you're reviewing any of these patches
15:21:50 <gouthamr> #link https://etherpad.opendev.org/p/manila-victoria-review-focus
15:22:23 <gouthamr> anything else we need to be looking at right away? apart from the specifications?
15:23:13 <gouthamr> alright, lets move on to the next slot
15:23:16 <gouthamr> #topic Bugs (vhari)
15:23:26 <gouthamr> vhari isn't around today
15:23:43 <gouthamr> but she's kept the etherpad ready for us
15:23:44 <gouthamr> #link https://etherpad.openstack.org/p/manila-bug-triage-pad-new (Bug Triage etherpad)
15:24:07 <gouthamr> #link https://bugs.launchpad.net/manila/+bug/1607150
15:24:07 <openstack> Launchpad bug 1607150 in OpenStack Shared File Systems Service (Manila) "Tempest test for dr/readable replication fails because share has two active replicas" [Medium,New] - Assigned to Douglas Viroel (dviroel)
15:24:43 <gouthamr> ah, dviroel picked this up recently - can we put this to victoria-3, dviroel ?
15:24:54 <dviroel> i think that this issue is still around, we might have one of our guys working on this
15:25:13 <dviroel> gouthamr: yeah, victoria-3 works
15:25:24 <gouthamr> dviroel: awesome, thanks
15:25:30 <gouthamr> #link https://bugs.launchpad.net/manila/+bug/1839640
15:25:30 <openstack> Launchpad bug 1839640 in OpenStack Shared File Systems Service (Manila) "NexenaStor5 NFS driver is not calculating provisioned_capacity_gb properly" [Medium,New] - Assigned to Alexey Khodos (alexey-khodos)
15:26:35 <gouthamr> hmmm, alexey reported this issue, and owned it - so i'll just check with him about the status
15:27:18 <gouthamr> its been around nearly a year too, so its likely not a medium priority as well - lets bump it down to low
15:27:36 <gouthamr> any other thoughts on this?
15:27:44 <gouthamr> #link https://bugs.launchpad.net/manila/+bug/1700660
15:27:44 <openstack> Launchpad bug 1700660 in OpenStack Shared File Systems Service (Manila) "DetachedInstanceError on trying to access share['share_server']" [Low,Triaged]
15:27:59 <gouthamr> oh boy this is an ooold and odd one
15:28:03 <tbarron> Are we still hittiing this issue?
15:29:00 <gouthamr> tbarron: i don't think we ever did - this was a tracker to investigate why drivers are hitting a DB error when the share server is "null"
15:29:26 <tbarron> by hitting it I men do drivers still get the DB error?
15:29:46 <gouthamr> nope, i think drivers handle it gracefully
15:30:21 <gouthamr> i think i was writing some new code when i hit this problem and wondered why we weren't returning primitives instead of handing db objects to the drivers
15:30:33 <tbarron> Seems to me like lazy load is fine but ^^
15:30:44 <gouthamr> dviroel carloss - does this ring a bell with any of your recent driver work?
15:31:03 <tbarron> manager shouldn't hand off db object to driver
15:31:22 <dviroel> gouthamr: we face some issues with db objects in the driver recently
15:31:39 <dviroel> gouthamr: when implementing create shares from snapshot in another pool
15:31:44 <carloss> dviroel there is a workaround in the code for this, right?
15:32:10 <dviroel> gouthamr: but we choose to convert to primitives before send to the driver
15:32:21 <tbarron> +1
15:32:25 <dviroel> this includes share_server info
15:32:26 <gouthamr> dviroel: oh great!
15:32:52 <dviroel> but now, share-manager has lot of db to primitive conversions
15:33:25 <carloss> an example https://github.com/openstack/manila/blob/e24a7402106f63e504cce053c0d2aff8f6709401/manila/share/manager.py#L4453
15:34:02 <gouthamr> yeah... that's undesirable; but a workaround for now
15:34:18 <gouthamr> we could experiment with setting expire_on_commit to True in the sessionmaker
15:34:37 <gouthamr> as noted in the bug...
15:35:06 <gouthamr> dviroel/carloss, could one of you update this bug with links to your recent work to workaround
15:35:42 <dviroel> gouthamr: I think that I can test your solution during this cycle, since we are implementing new driver's interfaces
15:35:51 <gouthamr> dviroel: awesome
15:36:02 <carloss> dviroel++
15:36:08 <gouthamr> we'll keep this around as a "low" and see if we can try things in the db layer
15:36:24 <gouthamr> thanks dviroel
15:36:31 <gouthamr> #link https://bugs.launchpad.net/manila/+bug/1651630
15:36:31 <openstack> Launchpad bug 1651630 in OpenStack Shared File Systems Service (Manila) "should use glanceclient instead of novaclient for image list" [Low,New] - Assigned to Zhenyu Zheng (zhengzhenyu)
15:37:22 <gouthamr> oh this is another old bug... we had a pretty good solution, but it broke the generic driver and we had to revert it
15:37:40 <gouthamr> i'll pop off the assignee
15:37:45 <tbarron> I'm trying to remember how it broke the generic driver
15:38:16 <gouthamr> tbarron: yes, logs are long gone, but we can revert the revert patch and see how the driver fares now
15:38:31 <tbarron> +1
15:38:41 <gouthamr> carloss - it's likely you'll hit this in your efforts around share server migration
15:39:09 <gouthamr> or maybe not, if you handle microversions carefully in the novaclient
15:39:36 <tbarron> we must have reverted under gate-breakage pressure and didn't revisit
15:39:39 <gouthamr> the thing is we *have* to make this change, or be stuck on a pre-2.36 microversion of nova
15:40:12 <gouthamr> tbarron: yes, looks like it was close to an RC timeframe
15:40:43 <gouthamr> any volunteers who'd like to investigate this?
15:40:43 <carloss> oh, that's true
15:41:22 <carloss> that'd happen only if we chose to improve the generic driver, right?
15:41:34 <carloss> s/chose/choose/
15:42:03 <gouthamr> carloss: choose to improve for the migration feature? yes
15:43:16 <carloss> yes, thanks, that's what I meant :)
15:43:31 <gouthamr> cool, we'll need some help with this tech-debt item - anyone here/reading - please assign it to yourself if you're willing to help!
15:44:09 <gouthamr> i'll post this discussion on the bug so we have context - likely the actions there are to revive that old patch, rebase and see what fails on the generic driver job
15:44:22 <gouthamr> #link https://bugs.launchpad.net/manila/+bug/1482633
15:44:22 <openstack> Launchpad bug 1482633 in OpenStack Shared File Systems Service (Manila) "requests to SSL wrapped sockets hang while reading using py3" [Low,Triaged] - Assigned to Victoria Martinez de la Cruz (vkmc)
15:45:24 <gouthamr> ah, this one is Invalid, vkmc?
15:45:31 <vkmc> I think so yes
15:46:22 <vkmc> let me double check, we haven't hit this because we are deploying with uwsgi by default
15:46:50 <gouthamr> vkmc: ack, and we've told folks to not use the old "manila-api" eventlet based wsgi interface
15:47:12 <vkmc> yep
15:47:21 <vkmc> gouthamr, do you consider we should debug and get this fixed though?
15:47:24 <gouthamr> vkmc: iiuc eventlet's ssl is still broken on python3.7
15:47:54 <vkmc> IIRC we were depending on eventlet folks to fix this... there were not much we could do
15:48:03 <vkmc> but situation might have changed
15:48:14 <gouthamr> vkmc: yes, i don't think its worth the investment - i know most folks are running manila-api with either uwsgi or mod-wsgi
15:48:28 <gouthamr> vkmc: lets mark this invalid, if there's someone concerned, they'll open another bug against manila
15:48:43 <vkmc> k
15:48:43 <tbarron> +1, if someone actually has this use case they should be complaining
15:48:55 <vkmc> I'll mark it as invalid
15:48:58 <gouthamr> vkmc: awesome, thank you!
15:49:20 <gouthamr> yay, we went through the whole thing, making vhari proud
15:49:29 <carloss> hehe
15:49:42 <gouthamr> #topic Open Discussion
15:50:19 <gouthamr> so this happened:
15:50:24 <gouthamr> #link http://lists.openstack.org/pipermail/openstack-discuss/2020-June/015638.html
15:50:49 <dviroel> awesome
15:51:16 <gouthamr> pretty cool, but neutron-dynamic-routing got new lieutenants - thanks tobberydberg_ and frickler!
15:51:39 <gouthamr> the project lives on...
15:52:01 <dviroel> \o/
15:53:32 <gouthamr> also thanks to tbarron for socializing our use case
15:54:01 <gouthamr> anything else to discuss today?
15:54:51 <gouthamr> ... going once...
15:55:24 <gouthamr> haha, you don't want the countdown
15:55:37 <gouthamr> thanks for attending everybody, keep reviewing! and stay safe
15:55:46 <tbarron> thanks all!
15:55:48 <dviroel> thanks
15:55:53 <gouthamr> #endmeeting