12:01:40 <ndevos> #startmeeting Weekly Gluster Community Meeting
12:01:40 <zodbot> Meeting started Wed Sep 16 12:01:40 2015 UTC.  The chair is ndevos. Information about MeetBot at http://wiki.debian.org/MeetBot.
12:01:40 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
12:01:51 <ndevos> #info Agenda: https://public.pad.fsfe.org/p/gluster-community-meetings
12:01:55 <ndevos> #topic Roll Call
12:02:11 <ndevos> hey all, who's there today?
12:02:15 * kkeithley is here
12:02:18 * rastar is here
12:02:20 * overclk is here
12:02:22 * msvbhat is present
12:02:25 * ira is here.
12:02:33 * raghu is jere
12:03:07 * poornimag is here
12:03:25 <ndevos> I have the feeling today could be quick, many developers in India seem to be taking todays afternoon and the next days off
12:03:53 * ira nods.
12:03:56 <ndevos> #info we're skipping action items assigned to people that are missing today
12:04:13 * tigert is here
12:04:19 <ndevos> #topic Last Weeks Action Items (assinged to people present)
12:04:24 <ndevos> #topic poornimag to send a mail on gluster-devel asking for volunteers to backport glfs_fini patches to release-3.5
12:04:48 <ndevos> poornimag: did you send the email out? I dont think I saw it
12:05:13 * rafi is here
12:05:44 <ndevos> poornimag: there are about 4 weeks until 3.5.7 will get released, maybe someone wants to send backports before that
12:06:00 * ndevos wonders where poornimag went...
12:06:14 <poornimag> ndevos, nope haven't sent it, will do it today
12:06:21 <ndevos> poornimag: okay, thanks!
12:06:28 <poornimag> ndevos, ok
12:06:35 <poornimag> ndevos, wc
12:06:51 <ndevos> #info poornimag will send the request for assistance on backports today
12:07:02 <ndevos> #topic rastar to initiate discussion on exploring the different approaches towards doing GlusterFS release management and announcements
12:07:23 <rastar> ndevos: haven't done that, but will do it today
12:07:23 <ndevos> rastar: could you elaborate the topic a little, we could not make much sense out of it last week
12:07:31 <rastar> have the draft ready
12:07:51 <rastar> oh, it is the same thing we discussed at the end of september 2nd gluster meeting
12:08:09 <rastar> on need to have a release planning/status page for all gluster releases/features
12:08:53 <ndevos> #info this topic is about the need to have a release planning/status page for all gluster releases/features
12:09:19 <ndevos> #info rastar will send an email to start the discussion later today
12:09:23 <ndevos> thanks rastar!
12:09:38 <ndevos> #topic closing old bugs on unsupported versions and removing unsupported versions from bugzilla
12:09:58 <ndevos> this topic came up during yesterdays bug triage meeting
12:10:17 <ndevos> obviously people are still opening new bugs against the 3.4 release that is End-Of-Life
12:10:34 <ndevos> we need to make clear to people that older releases are not supported anymore
12:10:47 <ndevos> but, we also would still like to give them the option to report bugs
12:11:12 <rafi> ndevos: +1
12:11:20 <ndevos> closing bugs against EOL'd releases is not very use friendly, and we're unsure if that should be done
12:11:27 <ndevos> s/use/user/
12:11:43 <kkeithley> maybe we should change, e.g., 3.4.7, to 3.4.7-EOL or something similar
12:11:58 <justinclift1> kkeithley: That sounds like a good diea
12:12:01 <justinclift1> idea even
12:12:40 <kkeithley> then when we triage we can set needinfo and ask if the bug still exists in newer versions?
12:12:40 <ndevos> yes, that, or create one (version-less) End-Of-Life version, and move all bugs to that
12:13:12 <kkeithley> if someone is looking for 3.4.7, will they even look for some generic EOL version?
12:13:18 <ndevos> personally I would be surprised if users understand EOL at all, and I would prefer to spell it out
12:13:31 <kkeithley> spelling it out is better, yes
12:13:58 <ndevos> I hope people would think "whats up?! my version is not listed" and check why that would be
12:14:18 <kkeithley> it'd be nice to think that. ;-)
12:14:31 <kkeithley> to think that that's what they'd do.
12:15:18 <justinclift1> It's probably a better idea to have the 3.4(.x?) bit included in the -End-Of-Life, so anyone doing metrics/stats can still work with the historical data
12:15:36 <justinclift1> If it all gets lumped into the same version, it sounds like that would be lost
12:16:12 <ndevos> not sure, there is a history tab in bugzilla that contains all those detail (no idea how to get that from a script)
12:16:21 <justinclift1> np
12:16:35 <ndevos> I also dont know who is doing statistics, if there is someone at all?
12:16:45 <kkeithley> maybe not now
12:16:54 <kkeithley> but some day?
12:16:58 <justinclift1> Yeah
12:17:13 <justinclift1> It's not an uncommon thought both inside RH and external
12:17:25 <justinclift1> People do all kinds of high level analysis of things
12:17:56 <ndevos> okay, so we'd like to rename the current EOL'd versions like: $version + "-End-Of-Life" ?
12:18:10 * justinclift1 nods
12:18:45 <ndevos> anyone else that does (not) like this?
12:18:47 * ira wonders if we are sure EOL is really... EOL? :)
12:18:52 <kkeithley> 3.4-end-of-life or 3.4.7-end-of-life?
12:19:13 * justinclift1 isn't sure
12:19:13 <ndevos> kkeithley: probably squash them all in one 3.4-end-of-life?
12:19:15 <kkeithley> ira: no, not if it's also a bug in later versions
12:19:36 <justinclift1> No preferences here
12:19:37 <ira> Then why the -eol tag?
12:19:37 <justinclift1> :)
12:19:45 <kkeithley> but the concern (or my concern) is that someone using 3.4.7 won't bother, or won't know if it's a bug in a later version
12:19:55 <ira> It's not a true eol, we haven't shot it dead, dead....
12:19:59 <ndevos> ira: well, 3.2, 3.3 and 3.4 are EOL, we will not update the git repo with fixes anymore
12:20:11 <kkeithley> then they file it against the only 3.4foo they can find, and we fix it in bug triage
12:20:47 <ira> Is there a note we can put in BZ when they file against it saying that is EOL?
12:20:56 <kkeithley> needinfo
12:21:24 <ira> No, when they select the version... so they get it instantly ;)
12:21:35 <ndevos> a note will not be automatic, we will do so during the weekly bug triage if the maintainers of the component fail to do it
12:21:37 <kkeithley> set needinfo and ask if they think it still exists in later versions. Or do our homework and check ourselves
12:22:41 <ira> In the last version announcement, you might wish to say, EOL.  I'm not sure  I'd put it in the version number.  I'd put it in the release notes, etc.
12:23:16 <ndevos> sure, its all in the release notes and such, but people still file new bugs against EOL'd releases
12:23:23 <kkeithley> We're talking about the Version: they can pick when they file a BZ
12:24:34 <kkeithley> we want to clear out all the BZs filed against EOL versions, and even delete the EOL versions from bugzilla
12:24:51 <kkeithley> for some definition of "we"
12:25:13 <ira> Do we allow versions 3.4.1 to be BZ'd?  If so they should get the -EOL tag also.
12:25:39 <ndevos> #info Fedora closes bugs that are filed against versions that are end-of-life
12:26:02 <justinclift1> Oops.
12:26:08 <justinclift1> AFK - Getting pizza out of ven
12:26:12 <ndevos> ira: yeah, there are a whole bunch of 3.4.* version in Bugzilla
12:26:13 <justinclift1> oven
12:26:38 * ndevos isnt a big fan of that, he rather sees one 3.4 version
12:27:05 <ira> I see both ways... But if they are all EOL... we should change it to be -EOL ;)
12:27:11 <kkeithley> close all 3.4.x BZs, delete all 3.4.x versions in bugzilla. Any new BZs for 3.4.x get filed against 3.4-end-of-life.
12:27:24 <ndevos> ira: yes, thats the plan :)
12:27:37 <kkeithley> Ditto for 3.3, 3.2, etc.
12:27:41 <ndevos> well, "plan", its the suggestions we're just discussing now
12:27:53 <rastar> kkeithley: but then we risk not knowing when the bug got introduced
12:28:05 <rastar> say a bug got in at 3.4.3
12:28:38 <justinclift1> Good point
12:28:41 <kkeithley> There's plenty of space elsewhere to document that sort of detail. right?
12:28:43 <ndevos> rastar: normally people post the exact version in the opening comment, the list of glusterfs packages installed
12:29:11 <kkeithley> the exact packages installed will have all that info as well.
12:29:26 <ndevos> or the logs...
12:29:52 <rastar> ndevos: they may skip unless it is necessary and I am just looking at it from search perspective
12:30:19 <rastar> version=3.4.3 is more accurate than description-contains:3.4.3
12:30:36 <ndevos> rastar: I doubt the real value it adds, IMHO the overhead of managing all those versions is worse
12:31:08 <ndevos> rastar: do you know of anyone using those details?
12:31:21 <rastar> Agreed. No automated way to reply to all new bugs filed with EOL message and automatically close it?
12:31:27 <kkeithley> rastar: I don't disagree, but....
12:31:48 <kkeithley> the value of diminishing returns....
12:32:15 <rastar> kkeithley: :), yes I agree, especially with the number of people who turn up for triage meeting
12:32:22 <ndevos> we could also just close all the EOL bugs with a note, and then remove the EOL versions from Bugzilla completelt
12:32:27 <ndevos> *completely
12:33:04 <rastar> Is the automated way bad?
12:33:30 <ndevos> rastar: if there is an automated way, it would not be too bad, but I dont think there is
12:34:04 <rastar> oh, I thought it would be similar to what we do for a new gerrit review message
12:34:26 <ndevos> no, thats quite different, Gerrit posts something in that case
12:34:50 <ndevos> we dont have a service/bot/.. that monitors for newly filed bugs and posts comments in the bugs
12:35:20 <rastar> ok, then it does make sense :)
12:36:04 <ndevos> I'll propose a vote, pick 'a' or 'b':
12:36:22 <ndevos> a. rename all 3.4.x versions to 3.4.x-end-of-life
12:36:45 <ndevos> b. close all EOL'd bugs with a note and remove the EOL'd versions from Bugzilla
12:37:00 * ndevos votes for B
12:37:58 <kkeithley> B includes creating, e.g., new 3.4-end-of-life versions in bugzilla
12:37:59 <kkeithley> ?
12:38:05 * kkeithley votes for B
12:38:13 * justinclift1 votes a
12:38:25 <ndevos> I would not even create a new end-of-life version, lets say thats optional
12:38:27 <justinclift1> Probably easier yeah?
12:38:38 <justinclift1> a) I mean
12:39:03 <ndevos> easier is not the question... what would be clearest for people that file new bugs?
12:39:03 * rastar votes for a
12:39:11 <rastar> tie breaker please
12:39:20 * justinclift1 still thinks A then
12:39:32 <rastar> rafi: ^^
12:40:03 <justinclift1> Lets ask on mailing list?
12:40:45 <ndevos> yeah, if we're tied we should get someone else to reply :)
12:40:53 <ndevos> kkeithley: could you send the email about that?
12:41:03 * raghu votes for b, but asking in the mailing list also looks fine
12:41:15 <kkeithley> I guess I don't actually care whether we keep eol BZs. If A is create, e.g., 3.4-end-of-life, reassign EOL BZs to 3.4-end-of-life, then delete all the 3.4.x versions in bugzilla.
12:41:33 <kkeithley> If that's it, then I change my vote to A
12:41:52 <kkeithley> yes, I will send email to the list
12:42:25 <ndevos> #action kkeithley will send an email to the list to get opinions on how to close/move EOL'd bugs
12:42:30 <ndevos> thanks kkeithley
12:42:36 <kkeithley> yw
12:42:37 <ndevos> #topic GlusterFS 3.7
12:42:49 <ndevos> hmm, no pranith...
12:43:00 <ndevos> anyone else who's watching progress of 3.7?
12:43:31 * ndevos guesses not
12:43:34 <ndevos> #topic GlusterFS 3.6
12:43:42 <ndevos> raghu: what are your plans?
12:43:57 <raghu> I planning to make 3.6.6 coming monday
12:44:08 <raghu> but unfortunately I dont have suffecients patches merged.
12:44:22 <ndevos> what is sufficient?
12:44:27 <raghu> After 3.6.5 only 3-4 patches have been merged
12:44:47 <justinclift1> New ETA?
12:44:49 <raghu> ndevos: If thats ok, then I am going to make 3.6.6 on monday
12:45:07 <ndevos> sure, thats fine
12:45:09 <raghu> otherwise I can wait for one more week, or I can make along with 3.7 release
12:45:26 <ndevos> #info raghu is planning to release 3.6.6 on Monday
12:45:41 <ndevos> nah, lets try to stick to the schedule a little
12:46:06 <ndevos> raghu: anything else?
12:46:15 <raghu> ndevos: there were some patches mentioned by richard from facebook, where he had backported some patches
12:46:41 <raghu> ndevos: but I think he mentioned only about commit ids, (especially in DHT and AFR)
12:46:51 <ndevos> raghu: oh, right, that would need some checking, I wonder if we already included some of the those
12:47:26 <raghu> ndevos: I dont think so. Because after his mail, I did not get any patches backported in AFR and DHT.
12:47:35 <raghu> probably I can consider them for 3.6.7
12:47:59 <ndevos> raghu: yeah, defer them to 3.6.7 if nobody has time before monday
12:48:20 <raghu> ndevos: sure. First I have to discuss about it with AFR and DHT folks
12:48:31 <raghu> 3.6.7 seems to be better for those patches
12:49:01 <ndevos> #info 3.6.x backport suggestions from Richard will not get included in 3.6.6, more likely for 3.6.7
12:49:24 <ndevos> thanks raghu!
12:49:35 <ndevos> #topic GlusterFS 3.5
12:50:04 <ndevos> I did a 3.5.6 release yesterday and sent the notification to the maintainers and packaging lists
12:50:23 <ndevos> A tracker 3.5.7 is available too
12:50:41 <ndevos> 3.5.6 includes 3 patches, two fixes and the release-notes :)
12:51:14 <ndevos> backports for 3.5 are much appreciated, please send them to Gerrit!
12:51:35 <ndevos> #info 3.5.6 has been released, announcement follows when more packages are available
12:51:44 <ndevos> #topic GlusterFS 3.8
12:52:13 <ndevos> ... I wonder who is here for 3.8 info
12:52:23 <ndevos> krishnan_p: maybe you?
12:52:56 <ndevos> #topic GlusterFS 4.0
12:53:12 <ndevos> probably the same, krishnan_p?
12:53:39 <overclk> ndevos, discussions have been going on for DHT2 with shyam
12:53:59 <overclk> ndevos, and for NSR with asengupt and jdarcy
12:54:09 <ndevos> overclk: ah, good, where did those discussions take place?
12:54:30 * ndevos did catch up on all emails to the lists yet
12:55:05 <overclk> ndevos, for dht2 as of now it's mostly b/w few folks and shyam takes care of putting things up on ML and gerrithub (as commits)
12:55:30 <overclk> ndevos (and all): https://review.gerrithub.io/#/q/project:ShyamsundarR/glusterfs
12:55:58 <ndevos> #info DHT2 progress can be followed on GerritHub https://review.gerrithub.io/#/q/project:ShyamsundarR/glusterfs
12:56:48 <ndevos> overclk: do you know if there is a DHT2 doc in the glusterfs-specs repository?
12:56:55 <overclk> ndevos, for NSR, the design document should be out for review shortly. jdarcy, any comments?
12:57:47 <jdarcy> Still trying to get caught up after being out of touch all day yesterday, but yeah, I think we'll be able to post a version of the spec pretty soon.
12:58:00 <overclk> ndevos, https://github.com/gluster/glusterfs-specs/blob/master/in_progress/dht-scalability.md
12:58:15 <ndevos> overclk: ah, cool
12:58:15 <jdarcy> Meanwhile, there's the six previous generations of specs out there, including one on Gerrit.
12:58:25 <overclk> ndevos, that's the closest I think in spec, but that's not what's going on in gerrithub ;)
12:58:37 * ndevos face palms
12:58:55 <overclk> ndevos, we surely need it to be updated.
12:59:18 <ndevos> overclk: is that something you can make happen? poke some people and get it updated?
12:59:25 <overclk> ndevos, I can probably take care of that
12:59:27 <overclk> ndevos, yup :)
12:59:51 <ndevos> #action overclk will get the dht-scalability doc in glusterfs-specs update to the latest design
13:00:27 <ndevos> #action jdarcy (and/or others) will post version of the NSR spec "pretty soon"
13:00:45 <ndevos> #topic Open Floor
13:00:47 <overclk> ndevos, and update glusterfs-spec too (for NSR).
13:01:04 <ndevos> overclk: yeah, I assumed that it ths spec that jdarcy mentioned
13:01:23 <ndevos> #info Weekly reminder to announce Gluster attendance of events: https://public.pad.fsfe.org/p/gluster-events
13:01:39 <ndevos> #info REMINDER to put (even minor) interesting topics on https://public.pad.fsfe.org/p/gluster-weekly-news
13:01:55 <ndevos> did anyone else have anything for the Open Floor?
13:02:52 <ndevos> well, if everyone stays quiet, we'll just close up for the day
13:03:00 <ndevos> thanks all for joining!
13:03:05 <overclk> thanks ndevos
13:03:11 <ndevos> #endmeeting