12:00:20 #startmeeting 12:00:20 Meeting started Wed Aug 26 12:00:20 2015 UTC. The chair is meghanam_. Information about MeetBot at http://wiki.debian.org/MeetBot. 12:00:20 Useful Commands: #action #agreed #halp #info #idea #link #topic. 12:00:39 #info agenda is available @ https://public.pad.fsfe.org/p/gluster-community-meetings 12:00:54 #topic roll call 12:01:05 * ndevos _o/ 12:01:25 * asengupt is present 12:01:28 * tigert is here 12:02:02 * saurabh_ saurabh_ is present 12:02:44 I'll wait for 2 more minutes before I move on to the next topic 12:02:46 * raghu is here 12:03:05 * kkeithley_blr is here 12:03:21 * krishnan_p is here 12:03:57 #topic Action items from last week 12:04:24 #topic raghu to fill in the contents for release schedule and ask tigert to push the page in gluster.org 12:04:51 Is raghu here? 12:04:53 meghanam_: oops. I have the contents. But I have to send the pull request to tigert. Will do it today 12:05:45 #action raghu to fill in the contents for release schedule and ask tigert to push the page in gluster.org 12:06:00 #topic msvbhat to do 3.7.3 announcement on the gluster blog and social media 12:06:26 * rastar is late for the meeting 12:06:28 * skoduri is here 12:06:48 Does anyone remember seeing this on social media? 12:06:58 yes 12:07:05 msvbhat: did that 12:07:05 rastar: ok 12:07:12 Cool. 12:07:21 erm, raghu even 12:07:29 * overclk is late 12:07:40 #info msvbhat and raghu announced 3.7.3 on social media 12:07:55 * anoopcs is late 12:07:59 meghanam_: here is the link https://medium.com/@msvbhat/gluster-news-of-the-week-30-2015-30452f44a144 12:08:12 #topic msvbhat/rtalur to send update mailing list with a DiSTAF how-to and start discussion on enhancements to DiSTAF. 12:08:17 * msvbhat joins the meeting bit late 12:08:18 * rjoseph is late 12:08:26 #info https://medium.com/@msvbhat/gluster-news-of-the-week-30-2015-30452f44a144 12:08:33 rastar? 12:08:48 That's the announcement, Yes 12:08:52 meghanam_: ^^ 12:09:04 Yes. I got that. Thanks msvbhat. 12:09:16 And about the distaf thing, I will create a review in gluster-specs and send for review 12:09:18 meghanam_: about DiSTAF, no announcement yet 12:09:19 DiSTAF topic. 12:09:22 as a proposal 12:09:28 had more discussion with ndevos 12:09:40 msvbhat: got an ETA for that doc? 12:09:57 ndevos: This weekend? 12:10:06 #info Discussion on DiSTAF enahancements started 12:10:28 msvbhat: thats up to you, asap :) 12:10:40 meghanam_: Create an action item on me to send a proposal request 12:10:47 in gluster-specs 12:11:11 #action msvbhat to send a proposal request for DiSTAF 12:11:12 msvbhat: I wont push the packaging too much until there is a clear user-facing stability 12:11:21 ndevos: I will try to complete tomorrow or day after. Weekend is my fallback if I couldn't finish then :P 12:11:44 ndevos: Sure 12:11:50 Okay. Moving on. 12:11:52 msvbhat: sounds good to me, people should then be able to review it starting next week 12:11:59 #topic kshlm to checkback with misc on the new jenkins slaves. 12:12:34 kslm, are you here? 12:12:44 kshlm? 12:12:56 or, maybe csim has something to report? 12:13:31 Looks like he is not here 12:13:38 #action kshlm to checkback with misc on the new jenkins slaves. 12:13:53 #topic poornimag needs help to backport glfs_fini patches to 3.5 12:14:09 poornimag, have you found any volunteers yet? 12:14:34 I did not see any patches flow in yet 12:15:15 Looks like she is not here too 12:15:40 #action poornimag needs help to backport glfs_fini patches to 3.5 12:16:03 #topic kslm to send out an announcement regarding the approaching deadline 12:16:03 * for Gluster 3.7 next week. 12:16:25 Does someone remember seeing this? 12:16:46 * kshlm is here 12:16:53 I'll send it out tonight. 12:17:00 #action kslm to send out an announcement regarding the approaching deadline for Gluster 3.7 12:17:11 And plan the release for monday 12:17:24 #topic raghu to make 3.6.5 this Friday. 12:17:29 Okay kshlm. 12:17:36 meghanam_: I made the release and the tarball is ready 12:17:48 raghu: what about various packages? 12:17:53 http://download.gluster.org/pub/gluster/glusterfs/3.6/LATEST/ 12:18:01 #info raghu released 3.6.5 and the tarball is ready. 12:18:07 nice, thanks raghu! 12:18:10 #info http://download.gluster.org/pub/gluster/glusterfs/3.6/LATEST/ 12:18:15 thanks raghu 12:18:28 hagarth: I think RPMs are ready. hchiramm was mentioning about Ubuntu packages to be prepared. I am waiting for that 12:18:31 raghu: announced on mailing lists? 12:18:37 will make the announcement once that is ready 12:18:54 raghu: ok, sounds good 12:18:58 #action raghu to announce 3.6.5 on mailing lists 12:19:00 hagarth: nope. Not yet. I can make it right now if RPMs are suffecient for that 12:19:06 I think the debian/ubuntu packaging was done today, kkeithley_blr? 12:19:37 ndevos: kkeithley did for one release of debian and one release of ubuntu today. 12:19:46 #topic hchiramm_ to improve release documentation 12:19:46 a few more to go 12:20:09 hchiramm_home? 12:20:22 Do you have any updates? 12:20:50 ubuntu was done today. debian wheezy is done. jessie and stretch will be done soon. I would not wait for those to announce 12:20:58 #action hchiramm_ to improve release documentation 12:21:02 kkeithley_blr: +1 to that 12:21:18 #topic Saravana_ to send out a mail to gluster-infra 12:21:38 an email about what? 12:21:40 I have sent out a mail to gluster-infra about that 12:21:50 FAQ on gluster-users mailing list 12:21:56 ah, right! 12:22:03 Alright. Thanks. 12:22:03 kkeithley_blr: shall I go ahead and announce 3.6.5 in the mailing list? 12:22:10 raghu: go ahead 12:22:23 #info Saravana_ sent FAQ on gluster-users mailing list 12:22:30 Saravana_: did you get any traction to that? 12:22:32 I need all your inputs on this. :) 12:22:33 raghu: yes, +1 to announce 12:22:52 #topic GlusterFS 3.7 12:22:52 where is the Ubuntu package link? I want to include that as well in the announcement 12:23:21 ubuntu dpkgs are in the launchpad.net/~gluster PPA 12:23:26 kshlm: you're up for the 3.7 release? 12:23:45 Yeah. 12:23:55 kshlm, ndevos I have couple of patches targetted for 3.7.4 and its not yet merged on master :( 12:24:07 I'll send out a mail announcing the approaching deadline. 12:24:08 kshlm, I have added the bug to 3.7.4 tracker, if that counts 12:24:12 kshlm, OK 12:24:25 krishnan_p: release was scheduled for later this week, would you be able to make that? 12:24:45 Looking at the 3.6.x release raghu is planning to do this week, I'm considering pushing 3.7.5 a little bit. 12:25:02 ndevos, I should be. This is the patch adding multiple threaded epoll to glusterd 12:25:03 * ndevos will be travelling tomorrow, and is unlikely in a position to review things before friday 12:25:23 ndevos, I have sent idea and few faqs(with answers) - prefer to have a link to gluster.org, Please get back with your suggestions. 12:25:24 I previously said, I was targetting monday, but I'd like to push it to next friday to get a better gap. 12:25:39 ndevos, ah! 12:25:42 krishnan_p: what necessitates multi-threaded epoll with glusterd? 12:25:43 kshlm: You can release the tarball as per ur schedule itself. By the time the packages are built, I think it would be right time 12:25:48 Does anyone have objections to this? 12:25:52 kshlm: agree with raghu 12:26:11 hagarth, to make snapshot commands work at 'limit 12:26:16 kshlm: if some patches miss this train, they will get picked up in the next month's release 12:26:19 krishnan_p: ah ok 12:26:19 without ping-timer between glusterd failing 12:26:36 kshlm, hagarth, ndevos it is not must-have 12:26:37 hagarth, raghu, from the last release experience I had, building the packages didn't take too long. 12:26:48 I'd like to have it. Will get necessary things going for it 12:26:49 * kkeithley_blr thought we were doing the jenkins release per the schedule, then the email announcement is made when "enough" packages land on d.g.o 12:27:09 and doesn't affect the schedule for the next 3.X release 12:27:23 kkeithley_blr: agree with that 12:27:48 kshlm: let us adhere to the announced release schedule. I think it will help us be more predictable. 12:27:55 kshlm: yeah. I think it takes atleast 1-2 days for all the packages to be available. And I think 1-2 days is ok. ??? 12:28:16 #info krishnan_p will try to get mulit-threaded epoll patches for glusterd ready before 3.7.4, but release will not be blocked if missing 12:28:30 Okay. 12:28:31 ndevos, the patches are already available for review 12:28:31 340858 12:28:53 krishnan_p: with "ready" is merged + backporting included ;-) 12:28:56 It has passed Linux regression, waiting for NetBSD to pass 12:29:02 Monday, is the deadline for 3.7.4 then. 12:29:03 ndevos, OK 12:29:09 kshlm, OK 12:29:15 kshlm: ok :) 12:29:34 #info Monday is the deadline for 3.7.4 12:29:34 kshlm: cool :) 12:29:48 #topic GlusterFS 3.6 12:30:06 raghu: ? 12:30:14 As I said, glusterfs-3.6.5 is released. I just have to make the announcement 12:30:17 kkeithley_blr: This is the link for Ubuntu packages right?? "https://launchpad.net/~gluster/+archive/ubuntu/glusterfs-3.6" 12:30:39 kkeithley_blr: shall I include that link in the announcement mail?? 12:30:47 raghu: yes. yes, include the link if you like 12:31:02 kkeithley_blr: sure. Thanks. 12:31:07 Moving on. 12:31:11 I will be sending out the announcement shortly 12:31:18 #topic GlusterFS 3.5 12:31:25 Thanks raghu. 12:31:31 poornimag? 12:31:35 Are you here? 12:31:51 bugs that need some work are listed on https://bugzilla.redhat.com/showdependencytree.cgi?maxdepth=2&hide_resolved=1&id=glusterfs-3.5.6 12:32:06 and the main one that gets user requests is the glfs_fini() cleanup 12:32:26 poornimag had sought help from volunteers last week 12:32:36 poornimag posted links to patches for backporting, and the backporting still needs to get done 12:32:50 Can someone help with the backports? 12:33:10 #halp Volunteers needed to backport patches listed in bug 1134050 12:33:35 #info bug with patches needed backporting: https://bugzilla.redhat.com/show_bug.cgi?id=1134050 12:33:56 #topic Glusterfs 3.8 12:34:29 hagarth: maybe you have something for 3.8? 12:34:34 or krishnan_p? 12:34:40 Are there any specific topics that needs discussion? 12:35:26 #info designs for 3.8 features should get posted to the glusterfs-specs repo REALLY SOON NOW 12:35:29 none from me 12:35:46 #topic Gluster 4.0 12:36:12 #info designs for 4.0 features should get posted to the glusterfs-specs repo pretty soon 12:36:20 agree with ndevos 12:36:42 all planned features for Gluster.next should start discussions upstream 12:37:10 We have shared our plan for glusterD 2.0 for the next 3 months. 12:37:15 and Gluster.Next is both 3.8 and 4.0 12:37:22 if you are working on a feature for gluster.next and have not reported progress in the last few weeks, please do so asap :) 12:37:31 We will be following up with emails discussing progress and open items 12:37:37 krishnan_p: a more granular progress update would be very welcome! 12:37:44 hagarth, definitely 12:37:50 krishnan_p: thanks for sharing the plan! 12:38:16 thanks. 12:38:22 #info discussions in Google Docs design documents need some summary on gluster-devel too 12:38:26 hagarth, ndevos would gluster weekly news be a good platform for sharing GLuster Next progress 12:38:36 krishnan_p: absolutely! 12:38:39 krishnan_p: sure, that would be great 12:39:05 hagarth, ndevos OK :) 12:39:18 #info Update Gluster News with Gluster.next feature progress on https://public.pad.fsfe.org/p/gluster-weekly-news 12:39:49 #action krishnan_p to update Gluster News about Gluster.next progress 12:40:01 #topic Open Floor 12:40:20 meghanam_, I think it is for everyone to update their respective progress in the etherpad so that weekly news can import it 12:40:28 krishnan_p: +1 12:40:54 how are we doing on regression failures? 12:41:06 Alright. Can you send an email to the devel list about this? 12:41:08 krishnan_p: yeah, that should work, but we're counting on you to show the way :D 12:41:35 ndevos, that is encouraging :) 12:41:42 krishnan_p: lead by example! 12:41:51 * krishnan_p ducks 12:42:08 Are there no existing topics for open floor? 12:42:11 should we enforce ban on patch acceptance for those components that cause regression failures? 12:42:15 hagarth: about regression , we might have fixed the NetBSD failures at last. Keeping fingers crossed. Lets see for a few days 12:42:28 rastar: cool, thanks for the update 12:42:29 hagarth: we have around 7 tests in bad_tests list 12:42:30 All, I recently bumped into a messaging library called nanomsg (nanomsg.org). 12:42:47 Would open floor be ideal for discussing about it and its relevance in gluster project? 12:42:54 krishnan_p, recently :D 12:42:54 tigert: I have created a pull request for the release-schedule to be updated in the gluster website. Please have a look at it and give me the feedback if there is any changes needed. 12:42:55 rastar: do all the tests in bad_tests have a bug associated with the,? 12:43:04 krishnan_p: sure 12:43:06 * msvbhat brb 12:43:11 kshlm, recently it all dawned on me. I have been looking at it for a while 12:43:16 ndevos: no but I agree with you, we should have one 12:43:30 ndevos: I will create them by end of this week. 12:43:42 rastar: yes, definitely, could you check and file bugs for the ones missing? 12:43:43 This library abstracts messaging for distributed systems software 12:43:55 tigert: this is the pull request (https://github.com/gluster/glusterweb/pull/5) 12:44:16 krishnan_p: what all messaging types does nanomsg support? 12:44:19 krishnan_p: looks interesting 12:44:26 At a high-level, it provides a socket like API to send/recv messages to a topology (read a group of network endpoints) 12:44:36 krishnan_p: looks interesting 12:45:04 hagarth, we need all endpoints use nanomsg to send/recv messages 12:45:21 krishnan_p: ok.. 12:45:22 hagarth, the application layer can have its own on-wire format 12:45:37 nice, just looked up the website 12:45:44 e.g, there is a topology called SURVEYOR 12:46:10 This has two endpoints, a surveyor (type socket) and one or more respondents 12:46:21 The surveyor binds on a well-known address 12:46:32 The respondents connect when they come 'online' 12:46:47 raghu: thanks 12:46:48 The surveyor sends a survey and sets a deadline for it 12:47:12 Performs as many recvs as surveyor deems fit (think quorum in case of consensus) 12:47:17 krishnan_p: I was wondering if there is a need for glusterd to speak to bricks? you would still need the current rpc/xdr too 12:47:31 Post deadline, messages from respondents are dropped 12:48:00 ndevos, there still is such a need. 12:48:11 krishnan_p: cool 12:48:29 This abstracts communication differently to our asynchronous send message over every connection 12:49:04 For instance, afr/dht could set a remote endpoint (topology) for all its bricks and issue the same message (when applicable) 12:49:23 nanomsg also supports publisher-subscriber pattern too 12:49:50 I would appreciate if interested folks hang out in gluster-dev and ping me with what you think :) 12:49:55 krishnan_p: seems very interesting 12:50:06 hagarth, I am a huge fan of it! 12:50:08 how does it communicate, does it use multicast or what? 12:50:16 thanks for sharing your discovery! :) 12:50:34 ndevos, the transport can be chosen. There is TCP, PGM and so on 12:50:47 So, if we chose TCP, there would be one-to-one message underneath. 12:50:51 krishnan_p: not only on #gluster-dev, send an email with the idea/suggestion to the list too 12:51:04 ndevos, I will. 12:51:32 ndevos, it is still nascent. I need some more collabortive thinking 12:51:33 #action krishnan_p to send an email about nanomsg.org to gluster-dev 12:51:55 * krishnan_p what is he getting himself into ... 12:52:13 krishnan_p: sure, and a pointer to nanomsg.org on the list will help in reviewing the options 12:52:25 ndevos, OK 12:52:41 Any other topics? 12:53:28 #info Weekly reminder to announce Gluster attendance of events: https://public.pad.fsfe.org/p/gluster-events 12:53:44 we are planning to decommission forge.gluster.org soon 12:53:48 #info REMINDER to put (even minor) interesting topics on https://public.pad.fsfe.org/p/gluster-weekly-news 12:53:54 if you have any concerns, please drop us a note 12:54:06 all projects on forge will mostly move to github's gluster account 12:54:24 #info forge.gluster.org will be decommisioned soon. 12:54:37 #info projects on forge.gluster.org will move to github.com/gluster/ 12:55:15 If there is nothing else, I'll end the meeting now 12:55:40 thank you meghanam_ and ndevos 12:55:41 Thanks to all the participants! See you next week :) 12:55:50 #endmeeting