15:06:19 <amye> #startmeeting 9-26-2018 Community Meeting
15:06:19 <zodbot> Meeting started Wed Sep 26 15:06:19 2018 UTC.
15:06:19 <zodbot> This meeting is logged and archived in a public location.
15:06:19 <zodbot> The chair is amye. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:06:19 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:06:19 <zodbot> The meeting name has been set to '9-26-2018_community_meeting'
15:06:42 <amye> Thank you bot. :D Amarts is the real host. :D
15:06:45 <amarts> hello all! Welcome you all to Gluster {User} Community meeting!
15:06:52 <amye> #topic 4.1.5 is announced
15:07:06 <amarts> all good with it, and packages available
15:07:13 <amarts> #topic triaging
15:07:39 <amarts> we did discuss about bug triage in previous maintainer's meeting
15:08:10 <sankarshan> what are the current challenges in addition to people's attention span?
15:08:23 <amarts> we want to have one of the maintainer take care of triaging on a round-robin basis, so our users can expect a response back from us in an week at the latest
15:09:02 * amarts any comments?
15:09:07 <sankarshan> oh! do you mean triage of incoming bugs?
15:09:13 <amarts> yes!
15:09:15 * kkeithley arrived late
15:09:21 <amye> hi kkeithley
15:09:35 <sankarshan> ok. how have the maintainers responded to this proposal?
15:10:22 <kkeithley> we used to have a weekly bug triage meeting. We stopped because maintainers promised to stay on top of the list
15:10:25 <amarts> the attendance in that meeting was from 8 people, and we _almost_ had agreement! First step is to get a set of bugzilla queries, so we are having proper set of things to look at
15:10:54 <amarts> kkeithley: true, as the promise was not very well maintained, we need to track it I guess
15:11:01 <kkeithley> and after we worked the list down to something maintainable
15:11:04 <sankarshan> alright. is someone creating the queries or, is that something on which help is being asked?
15:11:45 * sankarshan believes that the triage meeting kkeithley cited has queries and whines which can be reused
15:11:47 <amarts> nigelb and shyam volunteered to get the first step in this direction, and hopefully by next maintainer's meeting we will have some progress made
15:11:54 <sankarshan> splendid
15:12:08 <amarts> yes, we also get an weekly email on mailing list
15:12:10 <kkeithley> we already get a weekly email with the untriaged bugs
15:12:25 <kkeithley> {who,what}ever sends that email has the query.
15:12:33 <amarts> just want to make it more graphs with component based summary of bugs
15:13:21 <amarts> if there are no more questions on this, will move to next topic
15:13:42 <amarts> #topic release-5.0
15:14:01 <amarts> branching done! backports are being made!
15:14:09 <amarts> python3 is the *main* attraction
15:14:21 * shyam is here ...
15:14:22 <amarts> oh! the owner arrives
15:14:34 <shyam> Thanks to a ping from kkeithley :)
15:14:39 <kotreshhr> The components are not completely yet python3 compatible
15:14:53 <amarts> so i hear
15:15:06 <amarts> kotreshhr: which are these components?
15:15:19 <kotreshhr> I was looking into geo-rep.
15:15:25 * vbellur is here
15:16:23 <kotreshhr> strings needs handling. python3 stores it as unicode by default whereas in python2, it's bytes. It was giving problems
15:16:37 <amarts> thanks for this update!
15:16:41 <amye> cool, vbellur, we're looking for a GCS update after we get through triaging conversations.
15:16:48 <amarts> hopefully all of these would be addressed in next 4-5 days
15:16:59 * atinmu is here
15:17:12 <vbellur> amye: sure
15:17:21 <amarts> kotreshhr: while you are here, will the ctime feature be ready in release-5?
15:17:23 <kotreshhr> The challenge is to make it work both for python2 and python3. I will try to get those in
15:17:38 <amarts> i see a positive feedback from Anuradha and others in community after using it
15:18:17 <kotreshhr> ctime, yes. there is one pending patch for which shyam had queries
15:18:27 <amarts> ok, anyone else? got a chance to try out RC0 of release-5 ?
15:19:30 <shyam> Not yet, from my side. Will be getting to it today or tomorrow from a perf perspective
15:19:31 <amarts> ok, update for users, as we are taking this as notes!
15:19:47 <amarts> we have fixed many race conditions, which made our tests to fail
15:19:50 <atinmu> I see shyam mentioning some of the .t files failing, this needs some eyes
15:20:01 <vbellur> do we intend running any glusto tests on release-5?
15:20:04 <amarts> so recommended to update to release-5.0 after it is ready!
15:20:10 <amarts> loadtheacc: ^^
15:20:32 <amarts> atinmu: yes, we should
15:20:33 <loadtheacc> I would say yes.
15:20:39 <vbellur> it would be nice to have glusto tests passing as a pre-req for any release
15:21:06 <amarts> shyam: any help needed for the perf baseline tests?
15:21:18 <loadtheacc> vbellur, +1
15:21:21 <kkeithley> wrt 5.0rc0, unfortunately the release job whacked all the python3 shebangs back to python2
15:21:34 <amarts> vbellur: yes... i would say, we should get glusto to run nightly for sure, as a CI/CD pipeline in our upstream
15:21:44 <shyam> It would be nice yes, but we have not had it running or automated for the last 3-4 releases, nigelb is on it, but there are rough edges
15:21:44 <amarts> i guess loadtheacc is aware of the progress
15:21:57 <amarts> kkeithley: ok!
15:22:01 <amarts> :-(
15:22:07 <vbellur> ok, how do we make it happen?
15:22:22 <shyam> From a testing standpoint improvement nigelb has started nightly for regression and mux-regression from release-5
15:22:31 <sankarshan> i am not sure we've heard much about glusto runs happening for nightly builds or, what is required for that
15:22:51 <shyam> See: https://build.gluster.org/job/nightly-release-5/
15:23:03 <amarts> sankarshan: last i heard there were some issues of getting dependencies sorted
15:23:09 <amarts> shyam: thanks for the link
15:23:10 <loadtheacc> amarts, I've been focused on Python3 for Glusto, so will reach out to VijayA and Akarsha to include in an update.
15:23:23 <amarts> loadtheacc: that would be great!
15:23:46 <loadtheacc> amarts, VijayA will also be in the next maintainers meeting for an update
15:23:49 <vbellur> can we at least have a manual run before release-5? i.e. if automation needs more time?
15:24:49 * sankarshan would think that the issue here is about "who is going to own this item and drive it to completion"
15:25:10 <amarts> any volunteers? ^^^
15:25:42 <amarts> ok, any other topics on release-5.0?
15:25:42 <loadtheacc> Nigel has the keys to the Jenkins car, so we should pull him in.
15:25:57 <amarts> AI is we need an owner to run glusto-tests on release5-0 bits
15:26:49 <amarts> any one has questions?
15:26:52 <loadtheacc> amarts, I meet with VijayA tomorrow and will make this a topic to get a name
15:27:04 <shyam> I would not put a hard dep on "glusto on rel-5" as we dont know the effort. We need to get started, as a gate it would be good to have that for rel-6 (considering amount of time left on 5)
15:27:05 <amarts> loadtheacc: thanks, that would be very helpful
15:27:43 <amarts> shyam: agree! i don't think it would stop the release, but it would allow us to fix the known issues in 5.1
15:28:24 <amarts> #topic GCS
15:28:51 <vbellur> regarding GCS we are working on sending an update to the lists
15:29:06 <vbellur> atinmu: would you want to add some more thoughts here?
15:29:16 <atinmu> sure
15:29:56 <atinmu> As vbellur mentioned, community can expect an update on what has been achieved so far on the GCS umbrella pretty soon (mostly by next week)
15:30:20 * jstrunk *cheers*... We have a running (if fragile) cluster
15:30:21 <atinmu> we're also planning to start a weekly update from the individual projects falling under GCS to keep community up to date with the progress
15:30:36 <amye> that would be super, atinmu
15:31:05 <atinmu> GCS has its deployment script now (kshlm++) at https://github.com/gluster/gcs
15:31:16 * amarts says *cheers*!
15:31:20 <atinmu> which can be tested and feedback can be provided
15:31:34 <amarts> i see github issues started coming in
15:31:56 <atinmu> we're also working on the process part to ensure how do we track all these projects at single place, we're exploring tools like zenhub, waffle.io
15:32:20 <amarts> as a community, we need to have 'read' access to everyone?
15:32:25 <atinmu> at this point of time, me & jstrunk are in favor of using waffle.io over zenhub for its benefit
15:32:25 <vbellur> atinmu: awesome, will check it out
15:32:50 <atinmu> amarts, that was precisely one drawback which zenhub had, but waffle.io doesn't
15:33:21 <jstrunk> There may still be some rough edges wrt wafflebot and permissions, though
15:33:35 <atinmu> so yes, a log detailed email around all the activities of GCS is what is expected next week
15:33:53 <amarts> thanks for the update atinmu, jstrunk
15:33:59 <amarts> such an email would be great!
15:34:09 <amarts> ok, any questions on GCS?
15:34:59 <amye> Going once. :D
15:35:11 <amarts> quick update! Sadly, I have to run for another meeting!, so need help on following up below topics
15:35:17 <amarts> What is the best way to update community on progress across different project?Should we accept a project only if there is at least once in a month update on the project to mailing list?
15:35:24 <amarts> Should we continue to have the ‘releases’ note on this hackmd pad - amyeWe’re about to lose our chatlogs, botbot is going away in November.https://lincolnloop.com/blog/saying-goodbye-botbotme/
15:35:42 <amye> We should kill the 'release updates' here and move those to gluster-devel.
15:35:44 <amarts> need volunteers please
15:35:53 <amye> I'm on it, amarts.
15:36:00 <loadtheacc> Thanks, amarts!
15:36:00 <amarts> amye++
15:36:21 <amye> Additionally, the botbot.me service is going away November 5th.
15:36:25 <amarts> atinmu++ jstrunk++ loadtheacc++ shyam++ kkeithley++ kotreshhr++ sankarshan++
15:36:33 <amye> So if we want to keep logs around, we're going to have to find something else.
15:37:07 <amye> That was my last two things. Anything else?
15:37:39 <vbellur> amye: thanks for the heads up on botbot.me.. will be on the lookout for alternatives
15:37:54 <amye> I'll write a note to gluster-infra about it and see if good alternatives come up.
15:38:00 * shyam heading to another meeting...
15:38:02 <amye> We do have time, but it's going to be annoying.
15:38:30 <amye> If no other items of discussion, we can end this and give everyone ~20 minutes back.
15:38:43 <jstrunk> Do we need to discuss the "project update emails" item?
15:39:03 <vbellur> jstrunk: maybe park that for next time?
15:39:12 <jstrunk> fine w/ me
15:39:18 <vbellur> would be better to have amarts for that discussion
15:39:25 <amye> Yeah, I didn't think we had quorum for that.
15:39:47 <amye> Ok, next meeting is October 10th, any takers on hosting
15:40:44 * vbellur needs to run now. thank you all!
15:40:58 <sankarshan> 10th is it?
15:40:59 <atinmu> Even though it's not in today's agenda, but one of the item I need to bring up for next discussion is - how effectively we can manage that critical bug fixes aren't missed out in backports
15:41:00 <amye> Alright, I will host and call it a 30 minute meeting
15:41:04 <sankarshan> hmm i can give it a try
15:41:16 <amye> atinmu, edit the hackpad to add it and I'll put it in
15:41:28 <amye> sankarshan, save it for a week that I'm on the road. :D
15:41:29 <atinmu> amye, sure
15:41:51 <amye> Anything else in the last minute or so?
15:42:57 <amye> alrighty, good to see you all!
15:42:59 <amye> #endmeeting