Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2017-08-10

| Channels | #gluster-dev index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:13 bwerthmann joined #gluster-dev
00:48 baojg joined #gluster-dev
01:51 ilbot3 joined #gluster-dev
01:51 Topic for #gluster-dev is now Gluster Development Channel - https://www.gluster.org | For general chat go to #gluster | Patches - https://review.gluster.org/ | Channel Logs - https://botbot.me/freenode/gluster-dev/ & http://irclog.perlgeek.de/gluster-dev/
02:18 prasanth joined #gluster-dev
02:25 pkalever joined #gluster-dev
03:40 ppai joined #gluster-dev
03:44 riyas joined #gluster-dev
04:02 _ndevos joined #gluster-dev
04:02 _ndevos joined #gluster-dev
04:11 gyadav joined #gluster-dev
04:17 skumar joined #gluster-dev
04:43 atinm joined #gluster-dev
04:49 itisravi joined #gluster-dev
04:54 baojg joined #gluster-dev
04:54 ashiq joined #gluster-dev
04:57 nbalacha joined #gluster-dev
04:57 dkhandel joined #gluster-dev
04:58 rafi1 joined #gluster-dev
04:59 kdhananjay joined #gluster-dev
05:02 psony joined #gluster-dev
05:14 msvbhat joined #gluster-dev
05:15 ndarshan joined #gluster-dev
05:22 hgowtham joined #gluster-dev
05:22 gobindadas joined #gluster-dev
05:29 aravindavk joined #gluster-dev
05:39 Saravanakmr joined #gluster-dev
05:43 jiffin joined #gluster-dev
05:46 kotreshhr joined #gluster-dev
05:49 nbalacha joined #gluster-dev
05:55 sanoj joined #gluster-dev
06:05 apandey joined #gluster-dev
06:07 nbalacha joined #gluster-dev
06:18 ppai joined #gluster-dev
06:19 nbalacha joined #gluster-dev
06:20 hchiramm__ joined #gluster-dev
06:32 skoduri joined #gluster-dev
06:32 ankitr joined #gluster-dev
06:34 poornima_ joined #gluster-dev
06:38 ankitr joined #gluster-dev
06:41 msvbhat joined #gluster-dev
06:44 rraja joined #gluster-dev
06:46 Shu6h3ndu joined #gluster-dev
07:19 ankitr joined #gluster-dev
07:20 ankitr joined #gluster-dev
07:22 ankitr joined #gluster-dev
07:28 gyadav joined #gluster-dev
07:38 hgowtham joined #gluster-dev
07:42 msvbhat joined #gluster-dev
07:50 msvbhat joined #gluster-dev
08:01 nbalacha joined #gluster-dev
08:06 ppai joined #gluster-dev
08:09 hgowtham joined #gluster-dev
08:13 rastar joined #gluster-dev
09:36 rafi joined #gluster-dev
09:38 rafi1 joined #gluster-dev
09:41 nbalacha joined #gluster-dev
09:42 hgowtham nigelb, hi!
09:45 msvbhat joined #gluster-dev
09:49 riyas joined #gluster-dev
10:01 ankitr joined #gluster-dev
10:14 nbalacha joined #gluster-dev
10:21 skumar_ joined #gluster-dev
10:26 ndevos jiffin: what are the plans for the next (and last?) 3.8 update?
10:29 jiffin ndevos: I will be tonight 2mrw
10:29 kkeithley I appear to have missed the memo. Are we using cherry-pick now instead of submit in gerrit?
10:29 jiffin i will be travelling tonight
10:29 jiffin may be 2mrw
10:30 jiffin cli has one pending patch
10:30 ndevos kkeithley: the [cherry-pick] button can be used to backport the patch to an other branch, it will give you a pop-up before it posts a new change-review
10:31 jiffin atinm, samikshan: can anyone please ack this patch https://review.gluster.org/17766?
10:31 ndevos kkeithley: once you click the [submit] button, a comment will be added with "<someone> cherry-picked this change" or similar
10:32 bfoster joined #gluster-dev
10:33 kkeithley I've already posted the backports. They have +2. There is no submit button.
10:33 riyas joined #gluster-dev
10:33 kkeithley only cherrypick, rebase, abandon, and follow-up
10:34 nigelb kkeithley: link to patch?
10:34 kkeithley https://review.gluster.org/17927 is one of them
10:35 ndevos kkeithley: not everyone has permissions to merge patches in release-* branches anymore
10:36 kkeithley I see
10:36 nigelb kkeithley: Because you're not a release manager for 3.11.
10:36 nigelb Shyam and kshlm have permissions for that branch.
10:36 nigelb This changed from 3.10 onwards, I believe
10:37 nigelb (You probably didn't notice because you were one of the release managers for 3.10)
10:37 kkeithley nigelb: separate subject: did you see my question about debian packaging?
10:39 nigelb kkeithley: I did. I need to write to Patrick if he has any idea to deal with that gotcha.
10:39 nigelb I can't remember the right term to search for that.
10:39 kkeithley Patrick Matthai ?
10:40 kkeithley Matthaei?
10:41 kkeithley Or some other Patrick?
10:45 nigelb Same patrick.
10:53 nigelb kkeithley: I have a guess in terms of what to do.
11:00 baojg joined #gluster-dev
11:02 loadtheacc joined #gluster-dev
11:04 ndevos kkeithley: should nfs-ganesha be be part of gluster-3.12 repository for CentOS-6?
11:04 ilansch joined #gluster-dev
11:04 ilansch Hi all,
11:04 ilansch anyone alive ?
11:07 kkeithley ndevos: sure. why not?
11:08 ndevos kkeithley: I thought downstream would not have el6 from a certain release on, but I dont know the reasons for that - or I could be wrong completely
11:09 nigelb hgowtham: hey
11:09 hgowtham nigelb, I find the smoke tests are keeping on failing for two of my patches
11:09 kkeithley downstream doesn't. but downstream is irrelevant to here, right?
11:10 hgowtham https://build.gluster.org/job/smoke/36593/
11:10 hgowtham nigelb, ^^
11:10 ndevos kkeithley: I just wondered if there are technical restrictions with el6
11:10 hgowtham nigelb, the same error both the times. i didnt retrigger it again as i thought it might be something on the jenkins
11:11 ilansch I have a cluster with 2 nodes up and running, If from another "client machine"" i mount the gluster volume, mount -t nfs/glusterfs, when i execute operation from client machine, i listen to the operations in kernel vfs layer, if the client machine executed: touch somefile, then on the server machine, i get 3 events: 1.file create for "somefile" wit
11:11 ilansch h uid/gid=0/0, 2. file chown (uid to set is the user executed the operation from nfs client)) the user who executed the chown is 0/0, 3. file chgrp - same as #2..
11:11 kkeithley ndevos: nigling differences in pacemaker and the resources to test two distros are the reason nfs-ganesha isn't in rhgs-3.2(?, or rhgs-3.3 definitely)
11:12 ilansch is this the common flow gluster is working or i am doing something wrong ?
11:12 ilansch this is not a typical nfs flow.
11:13 ndevos kkeithley: ok, so just have everything for el6 too, thats fine for me :)
11:17 kkeithley If we had strong evidence that nobody is using nfs-ganesha on rhel6 then I'd say maybe we should drop it.
11:17 ndevos kkeithley: which version of nfs-ganesha + lintirpc do you expect in the gluster-3.12 repos?
11:18 kkeithley nfsg-ganesha-2.5.x and libntirpc-1.5.x would be my preference
11:18 ndevos ok!
11:19 kkeithley is that what we have in the gluster-3.11 repos too?
11:19 ndevos does anyone know what /usr/sbin/gluster-setgfid2path is abouf?
11:19 * kkeithley could look
11:20 ndevos nobody seems to care about 3.11, I had nobody confirming the test repos are working or not :-(
11:21 * kkeithley wonders why we bother with STM if nobody cares about them and nobody is using them
11:21 kkeithley <hansolo>I care</hansolo>
11:22 kkeithley I should update the Launchpad download counter to get numbers for 3.11
11:26 ndevos developers seem to care, the 6 months for releases were too long...
11:26 ndevos I would be really interested to hear results from users that tested 3.11 to give feedback to developers
11:28 kkeithley ndevos: I have some "smoke" (as in where there's smoke, there's fire) on three mem_get()s in dict.c for which there is never a mem_put().  If I start ganesha, mount, mkdir, copy file, unmount, stop ganesha there's  a set of mem_get w/ no mem_put.
11:29 kkeithley If I do the same but do two mkdir + two copy file, I get the same unmatched mem_get as the first run, plus another set of unmatched mem_gets
11:29 ndevos kkeithley: there are some dicts that contain options for the volume, or something, it would require unexporting the volume to see the dict_unref() and the mem_put() calls for those
11:30 ndevos not sure if there are other dicts that stay around for a long time
11:30 ndevos oh, well, if you stop nfs-ganesha, you'll probably find many other leaks too, some xlators dont clean up completely in fini()
11:31 kkeithley of course, but that's not the point.
11:31 ndevos it would be more interesting to start tracking mem_get() and mem_put() after starting nfs-ganesha, and the time before doing the mount, until after unmount, while ganesha stays running
11:33 ndevos if you repeat that a couple of times, subsequent runs should not really show more missing mem_put() calls
11:33 kkeithley The bonnie++ test doesn't unexport AFAIK, and if it's creating 1000s of directories during the run, and each mkdir leaks ~1K
11:33 glusterbot kkeithley: bonnie's karma is now 1
11:33 kkeithley bonnie--
11:33 glusterbot kkeithley: bonnie's karma is now 0
11:34 ndevos upon deletion of the directory, the caches for that dir should be free'd, so there should not be any leaks either
11:35 ndevos I guess that each mkdir adds something to the md-cache, and that could well be 1K per entry?
11:35 kkeithley I'm pretty sure these aren't mdcache
11:35 kkeithley but I need more data before I'm ready to say anything for certain.
11:36 ndevos yeah, getting the details is tricky
11:36 ndevos did you look at the statedump patches I emailed you? or tried them out?
11:37 kkeithley no, not yet
11:38 ndevos ok, np
11:42 msvbhat joined #gluster-dev
11:46 nigelb hgowtham: that was a problem with builder8.
11:46 nigelb I've taken it offline.
11:46 hgowtham nigelb, thanks ill retrigger
11:46 misc oh yeah, I was wondering why ansible failed for this one, but didn't had time to investigate :/
11:47 kkeithley nigelb: wrt debian packaging, if you have a guess what to do, please share.
11:54 kkeithley ndevos: maybe you know off the top of your head?  I have glusterfs-debuginfo installed. I get functions in libglusterfs in backtrace()+backtrace_symbols(), but no functions  in xlators ??
11:59 kkeithley oh, need to link with -rdynamic
12:00 kkeithley that would be a useful thing to have
12:09 kkeithley ndevos: and FWIW, the leaks I'm seeing appear to be coming from a dict in protocol/client, not mdcache
12:15 hchiramm__ joined #gluster-dev
12:20 baojg joined #gluster-dev
12:28 kotreshhr left #gluster-dev
12:37 misc nigelb: builder8 show as 100% cpu and nothing on the console
12:38 misc should I try to do some magic to diagnose the issue, or can i just restart it ?
12:40 Humble joined #gluster-dev
13:09 nbalacha joined #gluster-dev
13:32 Humble joined #gluster-dev
13:57 nbalacha joined #gluster-dev
14:11 aravindavk joined #gluster-dev
14:42 aravindavk joined #gluster-dev
15:10 wushudoin joined #gluster-dev
15:31 rafi joined #gluster-dev
15:41 baojg joined #gluster-dev
15:41 baojg joined #gluster-dev
15:42 baojg joined #gluster-dev
15:43 baojg joined #gluster-dev
15:44 baojg joined #gluster-dev
15:53 cholcombe joined #gluster-dev
16:09 aravindavk joined #gluster-dev
16:45 baojg joined #gluster-dev
16:57 vbellur1 joined #gluster-dev
16:59 vbellur joined #gluster-dev
17:02 skoduri joined #gluster-dev
17:33 msvbhat joined #gluster-dev
17:51 msvbhat joined #gluster-dev
18:10 vbellur joined #gluster-dev
18:22 ashiq joined #gluster-dev
18:29 ndevos kkeithley: these are the el6 packages for the gluster-3.12 repo: http://cbs.centos.org/koji/builds?tagID=1334
18:29 vbellur joined #gluster-dev
18:29 ndevos and http://cbs.centos.org/koji/builds?tagID=1330 for the el7 repo
18:31 ndevos once the sync is in place, the testing repos will be available on https://buildlogs.centos.org/centos/6/storage/x86_64/gluster-3.12 and https://buildlogs.centos.org/centos/7/storage/x86_64/gluster-3.12
18:31 humblec joined #gluster-dev
18:40 kkeithley ndevos++
18:40 glusterbot kkeithley: ndevos's karma is now 363
18:49 baojg joined #gluster-dev
19:45 bwerthmann joined #gluster-dev
19:51 baojg joined #gluster-dev
20:00 rafi joined #gluster-dev
20:22 msvbhat joined #gluster-dev
20:53 baojg joined #gluster-dev
21:32 vbellur joined #gluster-dev
21:54 baojg joined #gluster-dev
22:03 vbellur1 joined #gluster-dev
22:57 baojg joined #gluster-dev
23:00 vbellur joined #gluster-dev
23:58 baojg joined #gluster-dev

| Channels | #gluster-dev index | Today | | Search | Google Search | Plain-Text | summary