12:09:11 #startmeeting 12:09:11 Meeting started Wed Aug 13 12:09:11 2014 UTC. The chair is kkeithley. Information about MeetBot at http://wiki.debian.org/MeetBot. 12:09:11 Useful Commands: #action #agreed #halp #info #idea #link #topic. 12:09:16 #rollcall 12:09:32 * atinmu is here 12:09:35 kkeithley: uh, no, I have not seen that, just returned from lunch... 12:09:51 meeting minutes @ https://public.pad.fsfe.org/p/gluster-community-meetings 12:09:56 who's here? 12:09:58 * JustinClift is here, but really not in good state :( 12:09:58 roll call 12:10:11 * ndevos _o/ 12:10:40 * lalatenduM is here 12:11:16 * lalatenduM was expecting a lots of people to be here :( 12:11:19 well, that's sad 12:11:33 how many did the poll say, something close to 50 iirc? 12:11:33 let's carry on, we're already 10 min late 12:11:48 yes around 45 12:11:52 #topic action items from last week 12:12:39 Pranith to send a list of components and assignees? Did this happen? 12:12:56 afaik its not.. 12:13:03 I havent seen it... 12:13:16 #action pranithk to send a list with components and their assignees - soliciting others to contribute too 12:13:34 lalatenduM to send bug triage email? 12:13:41 working on it now I see 12:13:42 kkeithley, working on it 12:13:53 JustinClift to get initial GlusterFS Consultants and Support Companies page online 12:13:56 working on it 12:14:43 hagarth and pranith to communicate new plan for 3.6 on Aug 6th? 12:14:49 did that happen? 12:15:07 yes, they did 12:15:30 #action lalatenduM to send bug triage email 12:15:49 #action JustinClift low priority to get initial GlusterFS Consultants and Support Companies page online 12:16:03 ndevos to check on 3.6 nightly builds --- done 12:16:46 kkeithley & hagarth to follow up on fixes for 3.4.6 --- done. N.B. there are currently two low priority bugs in the 3.4.6 tracker 12:17:12 JustinClift to encourage Tamas Papp to log a bug bout  http://supercolony.gluster.org/pipermail/gluster-users/2014-July/018219.html 12:17:15 done? 12:17:20 Ugh 12:17:22 Totally forgot 12:17:33 #action JustinClift to encourage Tamas Papp to log a bug bout  http://supercolony.gluster.org/pipermail/gluster-users/2014-July/018219.html 12:17:37 tx :) 12:17:45 ndevos will verify status of patches proposed for 3.5.3 and request reviews for those patches in master? 12:18:02 ndevos, done? 12:18:10 ah, oh, no, not done, no ETA for 3.5.3 either 12:18:24 #action, no eta ndevos will verify status of patches proposed for 3.5.3 and request reviews for those patches in master? 12:18:42 JustinClift to create blog post of latest leaderboard stats, after receiving from hagarth 12:18:49 done? 12:18:58 Didn't receive from hagarth 12:19:06 So, blocked on that ;) 12:19:17 #action hagarth and JustinClift: blog post of latest leaderboard stats, after receiving from hagarth 12:19:25 Though, I could have done a better job of pinging him :) 12:19:35 Yeah, that AI works 12:19:36 kkeithley to email gluster-devel about the reviews needed for cppcheck stuff on master 12:19:52 sort of done, I nag, nobody reviews. :-( 12:20:22 sad kkeithley is sad 12:20:29 ;-( 12:20:39 kkeithley, I have reviewed these, but we need more experienced ones 12:20:46 yes 12:20:56 and btw, thanks for doing that. 12:21:11 lalatenduM: ^^^ 12:21:24 kkeithley, I believe I had also reviewed 12:21:34 atinmu++ 12:21:42 no karmabot here though. 12:21:56 :) 12:22:23 did we get all the AIs? the agenda is being edited in real time so I can't tell? 12:22:37 I just removed some duplicates 12:22:52 okay, then are we ready to move to the next topic? 12:23:01 I think so 12:23:06 #topic 3.4 12:23:43 tracker bug is https://bugzilla.redhat.com/show_bug.cgi?id=1125245, or https://bugzilla.redhat.com/show_bug.cgi?id=glusterfs-3.4.6 12:23:44 Bug 1125245: unspecified, unspecified, 3.4.6, kkeithle, ASSIGNED , GlusterFS 3.4.6 Tracker 12:23:45 Bug glusterfs: could not be retrieved: InvalidBugId 12:24:47 only two bugs currently. What does anyone think about a estimated date for doing a 3.4.6beta? 12:25:06 Early September perhaps? 12:25:16 +! 12:25:16 kkeithley, fine with me 12:25:18 +1 12:25:42 shall we say September 1st then? 12:25:44 well, one of them is urgent, do we know how often it happens? 12:26:46 #action JustinClift to test out Gitorious 3.x, to see if it's better than GitLab 12:26:51 (just putting this in there) 12:26:53 I'm looking to see if we have fixes for them? 12:26:58 both of them were reported by JoeJulian afaiu 12:27:12 I don't see a fix for https://bugzilla.redhat.com/show_bug.cgi?id=1123289 12:27:12 Bug 1123289: urgent, unspecified, ---, pkarampu, ASSIGNED , crash on fsync 12:27:13 no, I think not, or at least I dont see them - and pranith isnt here 12:27:52 until we get fixes we can't consider a date for a beta 12:28:02 yeah, thats what I was thinking 12:28:05 ndevos: http://review.gluster.org/8402 fixes the fsync crash I think 12:28:23 #action kkeithley will nag pranithk about fixes for 3.4.6 12:28:37 i don't think we have much information about the memory leak problem 12:28:53 hagarth: I think http://review.gluster.org/8296 should fix it :D 12:29:05 (the fsync one) 12:29:32 any other thoughts on 3.4? 12:29:45 ndevos: 8296 seems to be for a different problem ? 12:30:22 hagarth: the 3.4 bug is blocked by the bug that is used to post that change... (or am I looking at a different bz?) 12:30:42 hagarth: ah, yeah, never mind :-/ 12:30:51 ndevos: possibly, let us sync up on this one later. 12:31:07 hagarth: I'd leave that for kkeithley and pranith :) 12:31:27 ndevos: agree :) 12:31:48 * ndevos hands the token back to kkeithley 12:31:50 okay, already have an AI, we'll pull in hagarth if needs be 12:32:03 #topic 3.5 12:32:09 ndevos, the floor is yours 12:32:30 tracker with current bugs: https://bugzilla.redhat.com/showdependencytree.cgi?maxdepth=2&hide_resolved=1&id=glusterfs-3.5.3 12:33:05 lots of bugs IIRC. Are there fixes to go with them? 12:33:22 nothing extremely urgent, I think, so a release is not planned/dated yet 12:33:45 some have fixes, but the urgent ones not 12:33:56 well, not yet anyway 12:34:10 may I give you an AI to nag for the urgent fixes? 12:34:34 sure, go for it 12:34:47 #action ndevos to nag for 3.5.3 urgent fixes 12:34:48 ;-) 12:35:03 I guess it's more or less like: ndevos will verify status of patches proposed for 3.5.3 and request reviews for those patches in master 12:35:23 but yes, thats all the same in the end 12:35:29 I know you'll wait for fixes to be merged in 3.6 and master, but do we need to check if any need to be backported to 3.4? 12:36:23 hmm, maybe the DHT one, but you should have an email about it 12:36:39 kkeithley: yes, I would vote for the DHT fixes to be in 3.4.6 12:37:13 which bz is that? 12:37:53 bug 1116150 for master 12:37:54 Bug https://bugzilla.redhat.com:443/show_bug.cgi?id=1116150 high, unspecified, ---, nsathyan, POST , [DHT:REBALANCE]: Rebalance failures are seen with error message " remote operation failed: File exists" 12:38:04 and bug 1117851 12:38:05 Bug https://bugzilla.redhat.com:443/show_bug.cgi?id=1117851 high, unspecified, ---, jdarcy, POST , DHT :- data loss - file is missing on renaming same file from multiple client at same time 12:38:30 #action, kkeithley DHT backport 3.5.3->3.4.6 https://bugzilla.redhat.com:443/show_bug.cgi?id=1116150, https://bugzilla.redhat.com:443/show_bug.cgi?id=1117851 12:38:31 Bug 1116150: high, unspecified, ---, nsathyan, POST , [DHT:REBALANCE]: Rebalance failures are seen with error message " remote operation failed: File exists" 12:38:32 Bug 1117851: high, unspecified, ---, jdarcy, POST , DHT :- data loss - file is missing on renaming same file from multiple client at same time 12:38:40 anything else for 3.5? 12:39:01 not that I know of... 12:39:02 going once? 12:39:07 going twice? 12:39:14 #topic 3.6 12:39:24 hagarth, the floor is yours 12:39:41 3.6.0 tracker - https://bugzilla.redhat.com/showdependencytree.cgi?maxdepth=2&hide_resolved=1&id=glusterfs-3.6.0 12:40:06 hagarth: are all fixes required to be listed on that tracker? 12:40:21 we need a few fixes in there besides I am waiting for fixes from afr, dht and user serviceable snapshots 12:40:48 ndevos: not quite, need to be listed there if any bug/fix needs special attention 12:40:56 hagarth: ok 12:41:26 as per the 3.6 schedule, we intend doing a test weekend from tomorrow 12:41:33 hagarth: are you requesting cloned bugs from master -> 3.6.0 for patches that get posted, just like we do for 3.5 and 3.4? 12:41:50 +1 for cloning bugs 12:41:58 however I feel that we would still need at least the afr & dht fixes to conduct the test day 12:42:19 hagarth, +1 12:42:31 +1 for a test weekend - is there a test-plan? 12:42:44 ndevos: I am maintaining it lightweight right now. Probably we can bring in cloning after 3.6.0 is out? 12:43:13 ndevos: there are some test plans from 3.5.0. We need to provide the test weekend some structure too. 12:43:16 hagarth: no clones makes it difficult to close a bug when a release is made, we cant track two releases in one bug 12:44:29 We are putting that in to the triaging doc i.e. cloning bugs for each release 12:44:38 ndevos: ok, I am slightly bothered about increasing the overhead to get fixes into 3.6. However, if we think that cloning is helpful, I am all for it. 12:44:56 is there (general) reluctance to cloning because someone thinks a lot of open bugs "looks bad"? Or is it just the extra work to clone it and keep it updated? 12:45:14 kkeithley: It is the latter aspect 12:45:32 okay 12:45:50 if we get more effective at managing bugzillas, then it shouldn't become a problem. 12:45:54 I think the additional clicks to clone a bug have a greater benefit than messing up trackability of fixes 12:46:03 May be we should document it in patch backport wiki page 12:46:15 lalatenduM: +1 12:46:28 anyway, we need to resume our triaging discussions .. so let us figure out a way of managing bugs better. 12:46:52 sure, let us start cloning bugs for 3.6 too. 12:46:53 actually I think we should enforce it in the rh-bugs/smoke job, check the branch against the version the bug is filed 12:47:24 ndevos: that is a good idea if we can get to do it 12:47:26 ndevos: +1 12:47:37 hagarth: it;s not difficult :) 12:48:05 * ndevos wanted to propose that some weeks ago already, just seems to have forgotten about it 12:48:13 ndevos: I know, who owns that AI? ;) 12:48:20 ndevos, +1 12:48:30 #action: ndevos to add branch checking to rh-bugs 12:48:41 kkeithley: thanks :) 12:48:44 ;-) 12:48:50 #action ndevos to update the rh-bugs script in Jenkins to check git/branch and bug/version 12:48:52 coming back to the test weekend for 3.6 12:49:01 I propose that we move it by a week 12:49:21 to provide some structure to the test weekend and identify tests that we want to run 12:49:31 + 12:49:32 any opposed? 12:49:33 +1 12:49:40 hagarth, +1 12:49:57 #action hagarth to evolve plan for the test weekend and update schedule for 3.6 12:50:11 should we release alpha for test weekend? 12:50:18 lalatenduM: most certainly 12:50:33 ok :) 12:50:48 anything else on 3.6? 12:51:01 Hmmm 12:51:26 hagarth, one more point 12:51:29 As a data point, I've been working on getting cmockery2 formula into OSX Homebrew 12:51:33 atinmu: yes, please 12:51:42 When that's in, I'll do a GlusterFS one 12:51:50 hagarth, we are trying to make http://review.gluster.org/#/c/8126/ in for 3.6 12:51:51 JustinClift: cmockery2 is only for master atm? 12:52:18 Well, it seemed to be needed 12:52:20 hagarth, its still WIP though :( 12:52:42 atinmu: do you want to add the corresponding bug for this to the 3.6.0 tracker? 12:52:44 But yeah, the GlusterFS Homebrew formula can also pull master branch as well 12:53:00 So I've been making it work for both as much as possible :) 12:53:04 timecheck: eight minutes remaining 12:53:07 atinmu: it would be good to get this into 3.6.0 12:53:18 kkeithley: thanks, any other quick topics on 3.6? 12:53:34 hagarth, yes, I discussed this with Kaushal and we both agreed it should get into 3.6 12:53:43 atinmu: ok, sounds good. 12:53:56 kkeithley: don't seem to have more on 3.6. back to you again. 12:54:10 #topic: cppcheck fixes nag 12:54:12 'nuf said 12:54:25 #topic packaging 3.5.2 for Debian 12:54:46 did we decide anything last week about adding -dev and -geo-rep subpackages? 12:55:10 did you discuss with the debian maintainer? 12:55:17 I did not 12:55:34 maybe he should get involved in the decision making :) 12:55:45 #action kkeithley to discuss -dev and -geo-rep subpackages for Debian with pmatthai 12:56:22 #topic bugzilla sub-components for Community GlusterFS 12:57:20 are we interested in having, e.g., an xlator component with dht,afr,ec,etc.,etc., sub-components? 12:57:35 any other component/sub-components we would want? 12:57:40 we already have afr,dht, ec as components 12:58:03 cannot think of anything else atm 12:58:06 right, they would become xlator sub-components if we decide we want this 12:58:13 I think most important components are there, when we really do triage bugs, we'll figure out if we need others 12:58:29 hagarth, s/ec/disperse :) 12:58:44 Humble: ec is more appropriate for it though :) 12:59:11 +1 for ec 12:59:29 yeah , but we finalized on 'disperse' and its active . 12:59:33 :) 13:00:00 xavih, ^^^ 13:00:10 this came up with one of our program managers when I had him add new Version, Target-Release, and some other things, including some missing, new components. 13:00:49 anyway, it doesn't sound to me like there's a strong interest 13:00:52 kkeithley: ok, let us review and modify on a need basis 13:01:19 we're at one minute past the hour now. Shall we keep going? 13:01:20 kkeithley, I think with new component assignee task we can shrink or expand the list. thats what I think.. 13:01:30 kkeithley: let us quickly go over the rest ? 13:01:43 #topic web site mainenance 13:01:52 I added that topic 13:01:58 "404" errors-- 13:02:15 misc fixed it, and he's in #gluster-dev too now 13:02:19 maintenance 13:02:23 our website looks a lot better than before .. but we need to evolve a strategy for ongoing updates 13:02:30 +1 13:02:39 we need to evolve a proposal and document it on gluster.org 13:02:55 I think Eco wanted to write a doc for that... 13:02:56 who would be interested in evolving that proposal? I will start a discussion with the interested folks. 13:03:04 Count me in 13:03:24 misc will likely be an interested party as well 13:03:40 JustinClift: cool, that looks like a good list to start with :) 13:03:55 :) 13:03:57 #topic hagarth, JustinClift and misc to evolve a proposal for web site maintenance 13:04:00 oops 13:04:04 Heh 13:04:08 #action hagarth, JustinClift and misc to evolve a proposal for web site maintenance 13:04:24 #topic Where to propose and keep a list of projects for people who'd like to develop something? 13:04:30 who does this belong to? 13:04:34 feature pages on gluster.org 13:04:42 ah, yes, I added that with some ideas 13:04:49 would be the right place for such ideas 13:05:01 even if it is not a glusterfs feature? 13:05:18 and, how would interested people find it? 13:05:22 ndevos: as long as there is an integration possibility with glusterfs, we can have it on gluster.org 13:05:31 ndevos, may be a seperate page on gluster.org 13:05:46 ndevos: we can add a link from gluster.org->project proposals or something like that 13:06:03 Humble: hmm, yes, maybe in the 'how to contribute' and point to the proposed features 13:06:04 +1 13:06:39 okay, cool, let's figure out how to update the website first then 13:06:53 yep .. 13:06:59 and whenever there's a new feature page added, having a note on gluster-devel will ensure that it gets picked up for archival 13:07:42 yeah, and maybe ask gluster-users to propose their ideas too 13:08:10 i wonder if we need a better mechanism for managing feature pages but that's probably a topic for another day. 13:08:37 #action ndevos when we know how to update the webpage, add a pointer to proposed features for new contriburos 13:09:11 anything else (on this topic?) 13:09:21 no? 13:09:26 #topic Samba EPEL6 packaging issue http://supercolony.gluster.org/pipermail/gluster-users/2014-August/018374.html 13:09:51 #action kkeithley will talk to Ira and Jose about this. 13:10:26 kkeithley: cool 13:10:35 kkeithley, is the issue coming from samba pkgs in d.g.o 13:10:41 although our build w/ gluster support should be an exact clone of the Fedora/EPEL build 13:10:43 I believe so 13:10:59 * mikedep333 is here 13:11:09 because that's the only place you can get samba for epel w/ gluster 13:11:20 ah, yes 13:11:21 let me check if EPEL6 has samba packages e ven 13:11:35 *even 13:11:36 probably not 13:11:54 mikedep333, so you are using samba pkgs from d.g,o 13:12:01 d.g.o? 13:12:02 ? 13:12:03 it's not 13:12:06 download.gluster.org 13:12:09 yes 13:12:37 samba is not in EPEL because it's in RHEL 13:12:39 I assume that the Fedora packaging doesn't work correctly with EPEL6. It needs some update. 13:12:42 mikedep333, good catch thanks 13:12:42 right 13:12:57 Fedora packaging uses systemd. RHEL6 init.d 13:13:16 kkeithley, plz keep in me cc , I was thinking to add samba pkgs to CentOS SIG 13:13:23 But the .spec is 99% the same 13:13:35 I think the issue is where the PAM modules get installed 13:13:38 (and RHEL7 uses systemd) 13:13:41 right 13:13:42 the PAM modules have probably been moved 13:13:46 we'll get that sorted out 13:13:47 since RHEL6 13:13:49 great, thanks 13:14:28 I'll circle back with the our Samba guys who took over from me the builds for d.g.o 13:14:43 great :) 13:14:50 anything else on Samba? 13:15:05 1 second 13:15:11 * ndevos only wonders why we dont have only the vfs_glusterfs module 13:15:31 there was a samba 4.1.x security update 13:15:48 the d.g.o packages lack it I think. 13:16:09 http://www.samba.org/samba/history/samba-4.1.11.html 13:16:10 4.1.11, right? I think our builds lag a few days behind 13:16:14 yup 13:16:16 ok, understood 13:16:33 ah, and for such cases only the samba-glusterfs package would be nice :) 13:17:04 I'll bring that up with our Samba guys 13:17:29 any more (about Samba?) 13:17:46 #topic Backups for GlusterFS Infrastructure 13:17:59 JustinClift, that's you, right? 13:18:08 This is just me being worried, and wanting to find out if we have backups 13:18:24 Does anyone know of us doing backups for any part of our infrastructure? 13:18:29 JustinClift: right, let us discuss this on infra ML? 13:18:36 hagarth: Good idea 13:18:38 Lets do that 13:19:00 #action JustinClift to start discussion around backups on gluster-infra mailing list 13:19:03 #action JustinClift to discuss backup strategy on -infra ml 13:19:04 JustinClift, even the upgrades can be part of it :) 13:19:11 :) 13:19:22 okay, anything else? 13:19:31 going once? 13:19:45 going twice? 13:20:03 #endmeeting