15:01:32 <JustinClift> #startmeeting Weekly Gluster Community Meeting
15:01:32 <zodbot> Meeting started Wed Jul  9 15:01:32 2014 UTC.  The chair is JustinClift. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:32 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:01:45 <JustinClift> Etherpad URL: https://public.pad.fsfe.org/p/gluster-community-meetings
15:01:47 <glusterbot> Title: FSFE Etherpad: public instance (at public.pad.fsfe.org)
15:01:54 <JustinClift> Roll call.  Who's here?
15:01:59 * kkeithley is
15:02:01 * Eco_ present
15:02:06 * bennyturns is
15:02:41 * hchiramm__ here
15:03:17 <JustinClift> k, lets start
15:03:26 <JustinClift> #topic Action items from last meeting
15:03:27 <JustinClift> "JustinClift to follow up with Rodrigo Gonzalez about the Prominent Users page"
15:03:30 <JustinClift> Done
15:03:43 * hagarth is here
15:04:01 <JustinClift> Emailed Rodrigo abt 15 mins ago.  Cc-ing Eco.  We'll look to get this happening after the new website comes online
15:04:14 <JustinClift> "JustinClift to start the discussion on gluster-infra about updating links on the main gluster.org website. re 3.5.x links in template & the new Prominent Gluster Users page"
15:04:25 <JustinClift> Going to drop this.  Not needed any more with the new website soon happening
15:04:33 <JustinClift> "JustinClift to add freebsd slave node to Jenkins"
15:04:48 <JustinClift> Not done yet.  But it's possible to do now, since the FreeBSD stuff is merged
15:04:57 <JustinClift> Sometime in the next few weeks
15:05:11 <JustinClift> "hagarth to send out an email on abandon policy for gerrit patches"
15:05:12 <JustinClift> ?
15:05:30 <hagarth> JustinClift: haven't got to doing that, please carry forward that.
15:06:17 <JustinClift> np
15:06:34 <JustinClift> "hagarth will get initial functional leaderboard online"
15:06:39 <JustinClift> Initial version is online.
15:06:43 * JustinClift can't remember link atm
15:06:53 <JustinClift> hagarth: Does it auto-update atm?
15:07:08 <hagarth> http://employees.org/~vbellur/glusterfs/stats/patches-2014.txt
15:07:16 <hagarth> JustinClift: no, it is static.
15:07:35 <hagarth> I plan to keep updating that on 1st of every month
15:08:14 <JustinClift> No worries
15:08:31 <JustinClift> "JustinClift to get initial GlusterFS Consultants and Support Companies page online"
15:08:42 <JustinClift> Thinking to do this with the new website.
15:09:01 <JustinClift> get it ready now though
15:09:14 <JustinClift> "JustinClift to get autobuild test for NetBSD functional in Jenkins"
15:09:26 <JustinClift> We could turn it on atm.
15:09:46 <JustinClift> But thinking that's a bad idea, as we're near feature freeze cut off date and it could cause stuff to not get in
15:09:49 <JustinClift> I'd rather do it after feature freeze
15:10:03 <JustinClift> Ppl can update their NetBSD breakage then. ;)
15:10:09 <JustinClift> "JustinClift to look into the IPv6 restoring patch"
15:10:17 <JustinClift> Didn't get this done.  Probably won't either.  Working on completely different stuff atm.
15:10:32 <JustinClift> Does anyone else have an interest in trying out the IPv6 stuff, to ensure it works?
15:10:47 <JustinClift> ndevos: ^^^ ?
15:10:49 * xavih is here
15:11:04 <ndevos> JustinClift: interest yes, time no
15:11:44 <hagarth> ndevos: I have some thoughts, we can probably sync up later on IPv6
15:12:00 <JustinClift> Cool
15:12:11 <JustinClift> #action hagarth and ndevos to sync up on IPv6
15:12:16 <ndevos> hagarth: sure
15:12:27 <JustinClift> "pranithk & Nithya to write documentation for dht"
15:12:31 <JustinClift> That's done isn't it?
15:12:52 <JustinClift> Or still in progress?
15:12:53 <hagarth> JustinClift: thanks to jdarcy for this one!
15:12:59 <JustinClift> Ahhh, cool.
15:13:06 <JustinClift> I'll just mark it off as done. :)
15:13:18 <hagarth> work in progress, we should have it in soon with some minor edits.
15:13:24 <JustinClift> k
15:13:36 <JustinClift> "jdarcy to help pranithk and Nithya with their dht documentation"
15:13:37 <JustinClift> Done
15:13:42 <JustinClift> "JustinClift to redirect stderr lines for the regression testing (losetup -d) to /dev/null"
15:13:49 <JustinClift> Is this still actually a problem?
15:13:57 * JustinClift hasn't seen it recently
15:14:05 <JustinClift> But I haven't looked for it in last few days
15:14:33 <hagarth> JustinClift: I noticed it once over the last 2 weeks
15:15:12 <JustinClift> k.
15:15:15 <JustinClift> Non-urgent then
15:15:20 * JustinClift drops it
15:15:42 <JustinClift> "JustinClift to get new cleanup patch finished and merged"
15:15:55 <JustinClift> Not done.  Something in it is busting things.
15:16:09 * JustinClift needs to investigate properly (time consuming)
15:16:14 <JustinClift> So, "still in progress"
15:16:26 <JustinClift> "Pranith to find out if there's a way to run regression tests on a timed basis. Google, lpabon and bennyturns as info sources"
15:16:35 <JustinClift> pranithk: Howd' that go?
15:16:46 <pranithk> JustinClift: I think this was about long running jobs
15:17:04 <JustinClift> Yeah.  The status from the other week is "Complete - Rolling this out slowly.  Plan is to run FS sanity + perf regression daily, individual component automation weekly.  FS sanity is currently running, the rest will come as I get things stable."
15:17:05 <pranithk> JustinClift: Ben did find a bug also so I think it went well.
15:17:13 <JustinClift> Cool. :)
15:17:19 <JustinClift> It's still happening too?
15:17:21 <bennyturns> yep rolling things out slowly
15:17:34 <JustinClift> k, no worries
15:17:39 <JustinClift> Moving on
15:17:43 <bennyturns> all of my good HW is tied up at the moment
15:17:54 <JustinClift> bennyturns: Need moar budget
15:17:55 <JustinClift> :)
15:18:04 <bennyturns> :)
15:18:06 <JustinClift> #topic 3.4
15:18:16 <JustinClift> 3.4.5 beta 2
15:18:24 <JustinClift> kkeithley: ^
15:18:44 <hchiramm__> JustinClift, released and we are building  the rpms..
15:18:45 <JustinClift> How's this looking, or eta, or ?
15:18:50 <JustinClift> Cool. :)
15:19:23 <hchiramm__> :)
15:19:23 <JustinClift> hchiramm__: Would you be ok to do a blog post about 3.4.5 beta2 rpms being available, once they're ready?
15:19:29 <hchiramm__> sure .
15:19:44 <JustinClift> #action hchiramm__ to do a blog post about 3.4.5 beta2 rpms being available once they're ready
15:19:46 <kkeithley> yeah, what hchiramm__ said
15:20:05 <JustinClift> When do you guys reckon 3.4.5 release will be out, if beta2 seems good?
15:20:30 <hchiramm__> kkeithley, ur turn :)
15:20:32 <kkeithley> I'm on PTO for the next 7 business days (back June 20th). We can do the release then
15:20:48 <kkeithley> provided someone actually tests it
15:21:07 <kkeithley> ;-)
15:21:09 <JustinClift> :)
15:21:11 <hchiramm__> :)
15:21:24 <JustinClift> k, hchiramm__'s blog post should get people interested
15:21:48 <JustinClift> hchiramm__: Please ask for people to test it on the blog post.  Capital letters.  With blink tags type of thing
15:21:51 <hchiramm__> JustinClift, \o
15:21:53 <JustinClift> (ok, not blink tags) :)
15:22:02 <Eco_> blink tags ftw
15:22:17 <kkeithley> hypnotoad tags
15:22:35 * Eco_ is mesmerized
15:22:38 <JustinClift> ;)
15:22:41 <JustinClift> k, moving on
15:22:47 <JustinClift> #topic 3.5
15:22:58 <JustinClift> How're we looking for 3.5.2?
15:23:16 <ndevos> work in progress, patches get posted, reviewed and merged
15:23:18 <JustinClift> Do we have enough important stuff in it, or are we waiting for a few more issues to be resulved
15:23:29 * JustinClift is typing crappy today
15:23:36 <pranithk> JustinClift: I need to back port one mem-leak fix
15:23:42 <JustinClift> pranithk: np
15:23:55 <ndevos> some issues, nothing extremely urgent that I know
15:24:09 <ndevos> a 3.5.2 surely is coming, but ETA
15:24:14 <ndevos> unkown
15:24:17 <hagarth> ndevos: should we spin up a beta sometime?
15:24:21 <JustinClift> ndevos: Will we be able to put out a 3.5.2 beta1 after that mem-leak fix is in?
15:24:25 <pranithk> ndevos: I think we should roll out 3.5.2 beta after my fix?
15:24:47 <ndevos> hagarth: we can, but I'll have to check what other blockers were reported
15:24:47 <JustinClift> We can definitely get people for testing a 3.5.2 beta1 ;)
15:24:57 <pranithk> JustinClift: Who all?
15:25:07 <JustinClift> Corvid Tech will absolutely test it
15:25:11 <ndevos> #link https://bugzilla.redhat.com/showdependencytree.cgi?hide_resolved=1&id=glusterfs-3.5.2
15:25:12 <JustinClift> Other places will too
15:25:13 <glusterbot> Title: Dependency tree for Bug 1104511 (at bugzilla.redhat.com)
15:25:46 * ndevos *cough* seems people start to understand the blocker bug thing...
15:25:52 <JustinClift> Quite a few there
15:26:07 <JustinClift> ndevos: Yeah, that's good work.  Encouraging/repeating it until ppl get the hang of it
15:26:40 <JustinClift> We'll leave it as "still in progress".  No eta for beta1 yet.
15:26:52 * JustinClift hopes we can get a beta1 out in the next week
15:26:54 <hagarth> ndevos: should we send out an email reminder to folks who are working on fixes for 3.5.2?
15:26:54 <ndevos> I'll need to run my checker scripts on the bugs and move the ones with patches to POST, or MODIFIED etc...
15:27:06 <JustinClift> hagarth: Good idea
15:27:23 <ndevos> hagarth: when the bugs are in the correct state it makes sense, yes
15:27:37 <JustinClift> ndevos: Can you do that email out when it makes sense to?
15:28:02 <ndevos> JustinClift: sure, I think that part of the release-maintainers job :)
15:28:07 <JustinClift> ;)
15:28:24 <JustinClift> #action ndevos to remind people via gluster-devel to get their fixes into 3.5.2
15:28:34 <JustinClift> Anything else for 3.5.2?
15:28:43 <JustinClift> (or 3.5 at all)
15:28:44 <ndevos> not from me
15:28:49 <JustinClift> k, moving on
15:29:17 <JustinClift> #topic 3.6
15:29:26 <JustinClift> hagarth: How's this looking?
15:29:30 <ndevos> #link https://bugzilla.redhat.com/showdependencytree.cgi?hide_resolved=1&id=glusterfs-3.6.0
15:29:31 <glusterbot> Title: Dependency tree for Bug 1117822 (at bugzilla.redhat.com)
15:29:41 <hagarth> feature freeze for 3.6 is now on 07/12
15:29:47 <JustinClift> ndevos: Thanks
15:30:00 <hagarth> jiffin, anoopcs are working on trash and improved nfs configuration
15:30:03 <ndevos> hey, there is actually a bug on there already, I'm impressed
15:30:21 <hagarth> reviews for those by 12th will help quite a bit
15:30:47 <JustinClift> Anders added it.  He's pretty pro-active. :)
15:30:49 <ndevos> nfs configuration of nfs-ganesha?
15:31:04 <hagarth> as a matter of fact, reviews for anything and everything in gerrit will help :)
15:31:10 <JustinClift> Heh
15:31:16 <hagarth> ndevos: netgroups style configuration for gluster nfs
15:31:25 <JustinClift> After 3.6 is out, I'm thinking we have a focus on reviewing stuff that's hanging around
15:31:44 <JustinClift> Not do 3.7 / 4.0 until we've got that under control.  It should only take a few weeks.
15:31:49 <ndevos> hagarth: ah, right, I've seen some very early patches for that I think, but nothing recently...
15:31:56 <hagarth> ndevos: this was done by an intern in facebook last year, we are trying to get that into a glusterfs release.
15:32:06 <JustinClift> s/stuff that's hanging around/Gerrit CR's that are hanging around/
15:32:20 <hagarth> ndevos: right, anoopcs and jiffin will be sending out rebased patches soon
15:32:25 <ndevos> hagarth: yeah, I've sent jiffin an email about those patches a while back
15:32:46 <ndevos> (current NFS-patches for review: http://review.gluster.org/#/q/project:glusterfs+branch:master+status:open+message:nfs,n,z _
15:32:47 <glusterbot> Title: Gerrit Code Review (at review.gluster.org)
15:33:22 <hagarth> ndevos: cool, we probably should have one URL for all major components
15:33:23 <JustinClift> ndevos: Cool.  Didn't know it was possible to keyword search like that
15:33:23 <ndevos> JustinClift: sorry, dont want to get off-topic :)
15:33:30 <JustinClift> ndevos: np at all :)
15:33:31 <hagarth> any questions, thoughts on 3.6 ?
15:33:54 * JustinClift just hopes Kaushal's stuff on better peer id makes it in
15:34:11 <hagarth> JustinClift: we are close to getting that in
15:34:13 <JustinClift> So much future improvements can build on it
15:34:15 <JustinClift> Cool
15:34:25 <hagarth> and xavih's work on ec should also be merged tomorrow
15:34:46 <JustinClift> Do we know of stuff that's probably not going to make it?
15:35:12 <hagarth> so, we seem to have a nice set of features for 3.6 (snapshots, erasure coding, ssl improvements etc.)
15:35:18 <JustinClift> Cool
15:35:38 <JustinClift> k. moving on
15:35:46 <JustinClift> #topic Other Agenda items
15:35:53 <JustinClift> "rdma status update"
15:36:01 <JustinClift> kkeithley: Is that you?
15:36:03 <kkeithley> feel good news about RDMA
15:36:07 <kkeithley> yup, that's me
15:36:41 <kkeithley> I've used Mellanox's OFED drivers+packages on RHEL6.5 and Fedora19 (ubuntu, debian in the works, I don't have enough hardware)
15:37:17 <kkeithley> And using 3.5.1 I've got a pair of clusters with 6 DHT volumes, mounting native and nfs, and it all works
15:37:24 <JustinClift> Cool
15:37:24 <kkeithley> I'm going to run load later today
15:37:39 <JustinClift> Cool.  Done any perf testing?
15:37:41 <kkeithley> ubuntu and debian when I get back
15:37:47 <kkeithley> load/perf later today
15:37:48 <Eco_> kkeithley, do you still need to use IPOIB to set up the connection?
15:38:00 <kkeithley> yeah
15:38:14 <kkeithley> yes, I used IPoIB
15:38:27 <Eco_> for the entire test or just the initial connection?
15:38:50 <JustinClift> If it's transport == rdma, it'll just be for the initial connection won't it?
15:38:51 <kkeithley> just the initial connection. The volumes are "transport rdma" only
15:38:57 <JustinClift> Cool
15:39:01 <Eco_> cool
15:39:13 <Eco_> jinx you owe me a coke
15:39:13 * JustinClift hopes the perf tests aren't slower than non-rdma
15:39:21 <JustinClift> But, I guess we get to find out
15:39:28 <JustinClift> ;)
15:39:28 <kkeithley> do nfs is TCP on IPoIB, and the nfs servers are rdma to the gluster servers
15:39:43 <JustinClift> Makes sense
15:39:45 <kkeithley> s/do nfs/the nfs/
15:40:03 <JustinClift> k. moving
15:40:04 <JustinClift> on
15:40:09 <JustinClift> "trash translator"
15:40:13 <JustinClift> Who's is that?
15:40:28 <anoopcs> It's me
15:40:53 <JustinClift> k, go for it...
15:41:02 <jiffin> new patches uploaded
15:41:30 <jiffin> its working for deletion and internal operations
15:41:45 <hagarth> jiffin: by internal, do you mean unlinks due to self-heal and rebalance?
15:41:53 <jiffin> yup
15:42:05 <hagarth> jiffin: cool
15:42:25 <JustinClift> Reckon it'll make it into 3.6?
15:42:57 <anoopcs> I think we can make it into 3.6
15:43:05 <JustinClift> Cool. :)
15:43:13 <JustinClift> Good stuff.
15:43:23 <lalatenduM> jiffin, anoopcs nice!!
15:43:23 <jiffin> :)
15:43:25 <hchiramm__> anoopcs++ jiffin++
15:43:25 <hagarth> +1
15:43:31 <anoopcs> :-)
15:43:44 <JustinClift> k, are there any more agenda items, or do anoopcs or jiffin need more time for trash translator item?
15:44:23 <anoopcs> Any further clarifications on trash?
15:44:58 <JustinClift> Heh, the channel is yours.  Feel free to clarify if you want.
15:45:09 <hagarth> anoopcs: guess not, add lot more comments in code & add a test unit that tests everything in trash ;)
15:45:24 <JustinClift> Good thinking
15:45:43 <lalatenduM> hagarth++
15:45:47 <anoopcs> hagarth: I already wrote one trash.t file
15:45:54 <lalatenduM> also we need to run coverity with it
15:45:56 <ndevos> anoopcs: and a doc about the feature, explain how users can use it :)
15:46:06 <JustinClift> ndevos: Very good point
15:46:06 <anoopcs> ndevos: Sure
15:46:08 <hchiramm__> yep.. must required
15:46:15 <hagarth> anoopcs: cool
15:46:22 <lalatenduM> ndevos , nice point!
15:46:27 <JustinClift> k, thanks guys.
15:46:36 <JustinClift> Let's call it...
15:46:45 <hchiramm__> thanks :) \o/
15:46:48 <hchiramm__> bye
15:46:50 <JustinClift> #endmeeting