Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2014-10-28

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

All times shown according to UTC.

Time Nick Message
00:26 _Bryan_ joined #gluster-dev
00:27 shyam joined #gluster-dev
00:44 bala joined #gluster-dev
00:55 semiosis joined #gluster-dev
01:49 bala joined #gluster-dev
01:50 pranithk joined #gluster-dev
03:25 hagarth joined #gluster-dev
03:35 kshlm joined #gluster-dev
03:50 ira joined #gluster-dev
03:51 kshlm joined #gluster-dev
03:57 kanagaraj joined #gluster-dev
04:05 shubhendu joined #gluster-dev
04:23 atinmu joined #gluster-dev
04:37 rafi1 joined #gluster-dev
04:37 Rafi_kc joined #gluster-dev
04:39 anoopcs joined #gluster-dev
04:49 ppai joined #gluster-dev
04:55 ndarshan joined #gluster-dev
05:04 lalatenduM joined #gluster-dev
05:08 anoopcs1 joined #gluster-dev
05:16 soumya joined #gluster-dev
05:19 ira joined #gluster-dev
05:26 Humble joined #gluster-dev
05:27 nishanth joined #gluster-dev
05:29 deepakcs joined #gluster-dev
05:32 jiffin joined #gluster-dev
05:34 pranithk joined #gluster-dev
05:43 atinmu joined #gluster-dev
05:45 nishanth joined #gluster-dev
05:47 soumya joined #gluster-dev
05:50 atalur joined #gluster-dev
06:05 soumya joined #gluster-dev
06:19 nishanth joined #gluster-dev
06:26 anoopcs joined #gluster-dev
06:27 anoopcs joined #gluster-dev
06:29 ira joined #gluster-dev
06:30 ppai joined #gluster-dev
06:40 kshlm joined #gluster-dev
06:47 bharata-rao joined #gluster-dev
06:56 atinmu joined #gluster-dev
06:58 ppai joined #gluster-dev
07:25 rgustafs joined #gluster-dev
07:43 nishanth joined #gluster-dev
07:45 bala joined #gluster-dev
07:49 Humble joined #gluster-dev
08:43 hagarth ndevos: ping, what dependency does ganesha have on libgfapi? qemu seems to have it on libgfapi.so.0 and hence fails starting up
08:45 hagarth since we have libgfapi.so.7 now
08:45 kaushal_ joined #gluster-dev
08:56 hagarth xavih: should we mark quota and disperse as a known issue for 3.6.0?
08:59 xavih hagarth: I'm currently working on it. Anyway I'll need some time. Maybe it would be ok to to so. Do you plan to create another beta or the final release ?
09:00 xavih hagarth: there exists a simple workaround though, so it could be acceptable
09:05 hagarth xavih: I think we should document that workaround
09:05 hagarth xavih: I will probably do a beta4 today .. but we'll go GA sometime this week
09:06 xavih hagarth: ok. Where would be the best place to document this ? on the release notes ?
09:07 atinmu joined #gluster-dev
09:08 xavih hagarth: I feel that ec has not been tested much out there :(
09:12 hagarth joined #gluster-dev
09:12 hagarth xavih: yes
09:14 lalatenduM hagarth, the packages dependent on liggfapi.so.0 have to be rebuild for  libgfapi.so.7, we have plan to send an email to fedora devel mailing lists abt it once we plan for building 3.6.0 GA
09:14 ndevos hagarth: uhm, it depends, what nfs-ganesha are you looking at?
09:15 lalatenduM so that other package maintainers can rebuild their package if it is dependent on liggfapi.so
09:15 hagarth lalatenduM: ok
09:16 hagarth ndevos: nfs-ganesha in f20 or f21?
09:16 ndevos hagarth: oh, well, that will be the glusterfs version that ships with fedora, so glusterfs-3.5 and libgfapi.so.0
09:17 hagarth lalatenduM: thanks for that!
09:17 ndevos hagarth:as lalatenduM just mentioned, we can only update glusterfs in fedora when we announce that a rebuild of related packages is needed - so 3.6 can get inlcuded in f22 earliest
09:18 ndevos hagarth: but I think lalatenduM knows about a repository that has updated packages, like qemu and samba, probably nfs-ganesha should be part of that too
09:19 vikumar joined #gluster-dev
09:20 lalatenduM ndevos, nope, we plan to rebuild stuff for CentOS-SIG too, but after 3.6.0 GA happens
09:20 hagarth1 joined #gluster-dev
09:20 ndevos lalatenduM: ah, okay, so that is CentOS only, and not for Fedora too
09:21 hagarth ndevos: wouldn't we be pushing 3.6 earlier to fedora? pardon my ignorance wrt fedora packaging..
09:21 lalatenduM ndevos, once we send the mail to fedora devel and build 3.6.0 for rawhide, other maintainers will rebuild their stuff
09:22 deepakcs joined #gluster-dev
09:22 lalatenduM hagarth, we dont push beta builds to fedora , once it is GA , then we do
09:22 ndevos hagarth: no, we can not really do that - f21 is currently in beta and api-bumps like libgfapi.so.0 -> .7 are not really allowed
09:23 hagarth ndevos: ah ok
09:23 ndevos hagarth: we can try to get an exception from one of the Fedora boards, but it is something that is normally not done
09:23 hagarth lalatenduM: but in this case, it looks like we will need to wait for f22 to have 3.6 even after GA.
09:24 ndevos hagarth: the 3.6 will land in Fedora/Rawhide, and get included in the next fedora release
09:24 hagarth ndevos: right
09:25 ndevos we were just lucky that people did not really notice the 3.4 -> 3.5 updates in Fedora, there were no SONAME-bumps, so nothing needed to rebuild
09:25 lalatenduM hagarth, yes, we can go for an exception if we choose to i.e. we can add 3.6.0 in f21 too, but ideally on fedora version should not see a change in major version number in its life cycle
09:26 ndevos we can announce the 3.6 availability in Rawhide, and provide a copr.fedoraproject.org repository with rebuild packages for f20/21 if we really want to?
09:27 hagarth ndevos: yes, pushing 3.6 can be disruptive. are you aware of any practice that we can adopt to make it non disruptive if somebody were to install from Rawhide?
09:27 hagarth ndevos: that might be one possibility
09:28 ndevos I think that is the most common way for updated releases
09:28 hagarth ndevos: ok, maybe we should do that?
09:28 ndevos hagarth: we surely can, its very easy to setup
09:29 hagarth am pretty sure there will be considerable interest for trying out 3.6 given xavih's work, snapshots and other goodies in 3.6
09:29 ndevos yes, I guess so
09:30 hagarth ndevos: on that note, is there anything from ganesha's glusterfs fsal that we can highlight during GA announcement of 3.6.0?
09:30 ndevos lalatenduM: you want to setup a repo on copr.fedoraproject.org for 3.6 and rebuilds of related packages?
09:31 lalatenduM ndevos, I can do that. but I was thinking why not go for so version bump for f21 too?
09:31 ndevos hagarth: I was thinking about that, but the current work focusses on nfs-ganesha 2.2 and that is still in development
09:32 ndevos hagarth: maybe we can mention that there are many improvements/preparations for better integration with nfs-ganesha
09:32 hagarth ndevos: maybe we can update our howto on using nfs-ganesha 2.1 and glusterfs 3.6?
09:33 hagarth ndevos: yes, that would be good to have. would you mind updating that in minor features of release notes?
09:33 ndevos hagarth: yeah, we've been working on making things easier, I think an update on the howto would be good
09:33 ndevos hagarth: sure, I'll add that
09:34 hagarth ndevos: yeah, I am excited about all the good work happening there.. I think we can get some of the community interested too :)
09:34 hagarth ndevos++, thanks!
09:34 glusterbot hagarth: ndevos's karma is now 36
09:35 hagarth time for a karma check - @karma
09:35 hagarth @karma
09:35 glusterbot hagarth: Highest karma: "ndevos" (36), "lalatenduM" (29), and "JustinClift" (26). Lowest karma: "(" (-6), "(<" (-1), and "-rw-r--r" (-1). You (hagarth) are ranked 5 out of 45.
09:36 hagarth wow, we managed to get some negative karma for file access modes too ;)
09:40 ndevos hagarth: updated and draft sent to you :)
09:46 hagarth ndevos: merged, thanks!
09:49 hagarth ndevos: anything to add w.r.t gluster nfs in release notes?
09:50 ndevos hagarth: no, nothing I could think of
09:50 rgustafs joined #gluster-dev
09:51 hagarth ndevos: ok cool
09:52 hagarth ndevos: remember santosh adding wildcard support for auth-allow/reject in gnfs .. think we can add it in minor improvements?
09:54 ndevos hagarth: hmm, yes, if that is in 3.6 we can - I'll need to check that
09:54 hagarth ndevos: afair, it got merged before we branched
10:00 lalatenduM ndevos, have few questions on http://review.gluster.org/#/c/8418
10:01 ndevos lalatenduM: ask away!
10:02 kshlm joined #gluster-dev
10:03 lalatenduM ndevos, its abt %config(noreplace) %{_sysconfdir}/logrotate.d/glusterfs-georep , what is this mean
10:03 lalatenduM ?
10:03 ndevos lalatenduM: in case someone updated %{_sysconfdir}/logrotate.d/glusterfs-georep manually, the RPM will not replace the file
10:04 lalatenduM ndevos, cool ndevos++
10:04 glusterbot lalatenduM: ndevos's karma is now 37
10:05 lalatenduM ndevos, and this installs it from source , isn't it install -D -p -m 0644 extras/glusterfs-georep-logrotate \install -D -p -m 0644 extras/glusterfs-georep-logrotate \611
10:05 lalatenduM 628%{buildroot}%{_sysconfdir}​/logrotate.d/glusterfs-georep
10:07 ndevos lalatenduM: yes, that installs it into the buildroot
10:08 lalatenduM ndevos, ohh, which line installs it in in the OS?
10:11 ndevos lalatenduM: %{buildroot} should be the directory that the rpm packages? so when the rpm gets installed, it extracts the contents under /
10:13 atinmu joined #gluster-dev
10:14 ndevos ZFS--
10:14 glusterbot ndevos: ZFS's karma is now -1
10:16 hagarth ndevos: lol, how does zfs break health checker?
10:17 hagarth ndevos: nm, saw your response now
10:18 ndevos hagarth: uhm, yeah
10:19 ndevos if that is not sufficient, lalatenduM needs to write a patch that detects ZFS and does some other checking ;)
10:20 hagarth ndevos: yeah ;)
10:20 lalatenduM ndevos, ;), my ext4 patch is not excepted till now
10:21 hagarth lalatenduM: poke the maintainers :D
10:22 lalatenduM hagarth, :)
10:22 ndevos hagarth: http://review.gluster.org/8213 would be it
10:22 ndevos I'm always a bit hesitant to submit changes for components I'm not the maintainer of
10:23 ndevos and, I think others feel the same about that
10:24 hagarth ndevos: right, we probably should have ready to use queries for all our components
10:25 hagarth that way we can understand the backlog better and pay better attention to such components.
10:25 ndevos hagarth: sure, I have them for nfs and 3.5 :D
10:25 hagarth ndevos: yeah, I remember your query for nfs :D
10:26 hagarth ndevos: should we list all such queries in a wiki page or so?
10:26 ndevos hagarth: I guess we could
10:27 hagarth ndevos: yeah, let us see. if soumya deb can do some magic, we can actually make a cool dashboard with these queries.
10:27 ndevos hagarth: that would be awesome, a dashboard for patches and bugs would have my +1
10:28 hagarth ndevos: yeah, anything for better reporting!
10:36 ndevos hagarth: http://www.gluster.org/community/documenta​tion/index.php/Guidelines_For_Maintainers contains some of my bookmarks now, maybe you want to add yours?
10:38 hagarth ndevos: nice, will add mine there
10:39 ndevos hagarth: cool, thanks :)
10:40 hagarth lalatenduM: your patch LGTM
10:41 lalatenduM hagarth, ndevos is the patch fixes issue of Kiran Patil?
10:42 hagarth lalatenduM: not yet
10:44 lalatenduM hagarth, ndevos is "SPLError: 7869:0:(spl-err.c:67:vcmn_err()) WARNING: Pool 'zp2' has
10:44 lalatenduM encountered an uncorrectable I/O failure and has been suspended" is zfs specific error?
10:44 lalatenduM Copied the error from the mail thread
10:47 lalatenduM ndevos, hagarth, ok it seems the hung is before the brick failure detection
10:47 lalatenduM for zfs
10:49 nishanth joined #gluster-dev
10:53 ndevos lalatenduM: the "has been suspended" is probably key here, suspended may block all I/O, and not return any errors
10:54 ndevos lalatenduM: I think it is similar to running the fsfreeze command on the filesystem where the brick is located, creating a lvm-snapshot would do the sane
10:54 ndevos *same
10:54 badone joined #gluster-dev
11:00 lpabon joined #gluster-dev
11:02 ndevos lalatenduM: oh, now you can also practise the backport guidelines for backporting your patch to 3.5 and 3.6 :)
11:04 ppai joined #gluster-dev
11:05 hagarth hmm, regression runs don't seem to be automagically happening on release-3.6 patches
11:05 ndevos JustinClift: regression tests are piling up again, could you fix some of the offline slaves?
11:06 ndevos hagarth: yes they do, but there is a queue
11:07 ndevos http://build.gluster.org/job/rac​kspace-regression-2GB-triggered/ contains some release-3.6 patches
11:07 hagarth ndevos: ah ok
11:08 ndevos oh, and the last regression tests seem to take almost 2 hours?
11:10 hagarth ndevos: that's surprising .. worth a check why it's taking longer
11:13 ndevos hagarth: maybe it has to do with the failures, but I do not see anything obvious immediately: http://build.gluster.org/job/rackspace​-regression-2GB-triggered/2243/console
11:14 ndevos and http://build.gluster.org/job/rackspace​-regression-2GB-triggered/2241/console was not faster and did not hit an error
11:15 hagarth yeah .. the time spent for running each test unit in 2243 doesn't seem to be significantly bad
11:16 ndevos there are some with 100+ seconds, but that may well be needed, I guess we just need more builders online
11:18 hagarth ndevos: right
11:20 lalatenduM ndevos, yw! yup, will send back port :)
11:22 ws2k3 guys do you know if there are nagios plugins for glusterfs Without using nrpe?
11:26 hagarth kanagaraj: any help for ws2k3's question above?
11:27 kanagaraj ws2k3, all the active check plugins uses nrpe
11:27 kanagaraj ws2k3, but there are some passive check plugins which don't use nrpe
11:29 kkeithley (05:15:59 PM) lalatenduM: kkeithley, regarding http://review.gluster.org/#/c/8418
11:29 kanagaraj ws2k3, passive check plugins monitor the status of various services/processes like glusterd, nfs, smb, shd and server quorum
11:29 bala joined #gluster-dev
11:29 kkeithley JustinClift: I think it was jclift, let me double check
11:29 kkeithley lalatenduM: ^^^^
11:29 lalatenduM kkeithley, got it answered from ndevos :), will givemy +1 to it
11:30 kanagaraj ws2k3, these plugins uses nsca to send notification to nagios server
11:30 nishanth joined #gluster-dev
11:31 lalatenduM kkeithley, ndevos last question , so %{buildroot} actually becomes the right path for a RPM when it is getting installed
11:31 lalatenduM thats what I getting from http://fedoraproject.org/wiki​/Packaging:Guidelines#Using_.25.7Bbuildroot.7D_and_.25.7Boptflags.7D_​vs_.24RPM_BUILD_ROOT_and_.24RPM_OPT_FLAGS
11:36 hchiramm joined #gluster-dev
11:38 lalatenduM kkeithley, ohh ndevos has already cherry picked it
11:42 ws2k3 so why i would need nrpe for active checks
11:42 kkeithley %{buildroot} is ~/rpmbuild/BUILDROOT/....    That's essentially a staging directory for assembling the RPMs
11:43 ws2k3 correct me if i'm wrong but if there would be an application that can check all the bricks etc that could be done over tcp
11:44 ws2k3 the issue is kinda the gluster plugin i use now does give me a critical from time to time while i am sure the gluster is working fine and all brick are available
11:45 kkeithley lalatenduM: ^^^
11:45 lalatenduM kkeithley, when we install a RPM what macro we should use to place a binary in right path
11:46 soumya_ joined #gluster-dev
11:46 kanagaraj ws2k3, nrpe works like an agent for nagios server
11:47 kanagaraj ws2k3, we could use ssh instead, but nrpe provides more control over the execution
11:48 kanagaraj ws2k3, if you could send me the steps which leads to frequent critical alert, i can look at it
11:51 kanagaraj ws2k3, also are you getting critical alerts for bricks or some other services?
11:52 ws2k3 hmm i now noticed that the server has a high WA
11:54 kanagaraj ws2k3, it a particular plugin takes longer time to execute, nrpe might timeout and hence the alert is generated
11:55 ndevos REMINDER: Gluster Bug Triage meeting in #gluster-meeting starts in 5 minutes
11:56 ws2k3 kanagaraj it looks like glusterfs is eating up the IO for the OS disk
12:01 hchiramm_ joined #gluster-dev
12:02 ws2k3 the servers that i am currently using as glusterfs servers the OS is installed on a USB drive is it possible that that is causing issues ?
12:06 soumya joined #gluster-dev
12:18 kkeithley lalatenduM: %{buildroot} is used in the .spec file, not (AFAIK) when you install an RPM.  E.g. to install a systemd .service file you'd have a line like
12:18 kkeithley install -D -p -m 0644 src/scripts/systemd/nfs-ganesha.service %{buildroot}%{_unitdir}/nfs-ganesha.service
12:20 kkeithley during the rpmbuild, nfs-ganesha.service is installed into the staging directory ~/rpmbuild/BUILDROOT/nfs-ganesh​a-2.1..../usr/lib/systemd/....
12:20 kkeithley when the rpm is installed, nfs-ganesha.service is unpacked into /usr/lib/systemd/....
12:21 kkeithley you can build rpms to be "relocatable."  If you do that then you can do some magic command line option that result in things being installed in, e.g., /opt/foo/bar/usr/lib/systemd/...
12:22 kkeithley does that answer your question?
12:23 kdhananjay joined #gluster-dev
12:24 lalatenduM kkeithley, yes, got it thanks kkeithley++
12:24 glusterbot lalatenduM: kkeithley's karma is now 26
12:25 bala joined #gluster-dev
12:46 ndevos hagarth: just wondering if you have seen https://bugzilla.redhat.co​m/show_bug.cgi?id=1147236 before?
12:46 glusterbot Bug 1147236: high, unspecified, ---, gluster-bugs, NEW , gluster 3.6.0 compatibility issue with gluster 3.3
12:51 rgustafs joined #gluster-dev
13:01 jdarcy joined #gluster-dev
13:09 hagarth ndevos: should we have 3.6 compatible with 3.3?
13:11 kshlm joined #gluster-dev
13:19 ilbot3 joined #gluster-dev
13:19 Topic for #gluster-dev is now Gluster Development Channel - http://gluster.org | For general chat go to #gluster | Patches - http://review.gluster.org/ | Channel Logs - https://botbot.me/freenode/gluster-dev/ & http://irclog.perlgeek.de/gluster-dev/
13:22 hagarth ndevos: 3.5 is compatible with 3.3 afaiu .. I have tried 3.4 and 3.3, 3.5 and 3.4 deployments .. they work fine
13:23 hagarth ndevos: sorry about today's meeting, I thought that slot was available and scheduled a different meeting then
13:24 ndevos hagarth: ah, good to have that confirmation :)
13:24 hagarth ndevos: did your TZ have any DST changes recently?
13:24 * ndevos goes for lunch, bbl
13:24 ndevos hagarth: yeah, last weekend
13:24 hagarth ndevos: that explains it
13:25 ndevos hagarth: next week it's the US's turn, things will get messy even more
13:25 hagarth ndevos: probably, we will need to re-send the invite but let us think about that after your lunch :)
13:25 ndevos hagarth: yeah, I can do that later
13:25 * ndevos leaves before he's late
13:34 bala joined #gluster-dev
13:34 johnmark show of hands - who's going to openstack summit?
13:35 johnmark any core committers?
13:52 jobewan joined #gluster-dev
14:23 wushudoin joined #gluster-dev
14:43 tdasilva joined #gluster-dev
15:21 xavih this regression test seems to be stuck and blocking smoke tests: http://build.gluster.org/j​ob/regression/4932/console
15:30 ndevos xavih: oh, right
15:31 ndevos xavih: I've aborted it now, and will send an email to whoever started a regression test there
15:38 hagarth johnmark: deepak and csaba would be at the summit
15:44 johnmark hagarth: ok, cool
15:44 ndevos hagarth: oh, cool, they are attending the Gluster Community night too?
15:45 hagarth ndevos: we can certainly notify them.. we should create a blog post on the event
15:45 ndevos hagarth: yes, we should!
15:46 ndevos davemc: will you blog about the Gluster Community night that happens next week?
15:50 davemc ndevos, as soon as I can confirm we have an event.
15:52 ndevos davemc: ah, right
15:53 johnmark davemc: that's kind of important :)
15:55 davemc Are we on schedule to release 3.6.0 this week? We want to do a quick video of features if so
15:55 davemc (we being Brian Chang and me)
15:56 hagarth davemc: struggling hard to get it out.. we can do a sanity check in tomorrow's meeting
15:57 ndevos lalatenduM: still there? how is that logrotate fix going?
15:58 ndevos lalatenduM: if there is no response from you soon, I assume you're enjoying your evening and I guess I'll send a fix out later today
15:58 hagarth davemc: I am inclined to get it out this week.
16:01 lalatenduM ndevos, working in it , i had to understand how it works, then trying to fix it :)
16:02 lalatenduM ndevos, did you see back ports of beick failure detection for 3.5 and 3.6
16:08 ndevos lalatenduM: cool, thanks!
16:08 ndevos lalatenduM: I've seen updates in the bugs about those, I'll review them in a bit
16:12 lalatenduM ndevos, logging off now, will login back from home
16:26 ndevos hagarth: I've prepared a new Fedora COPR repository where I can trigger rebuilds for related packages, I'll test with a 3.6-beta tomorrow
16:26 ndevos http://copr.fedoraproject.org/co​prs/devos/glusterfs-3.6-stable/ would be the link
16:27 hagarth ndevos: awesome!
16:27 hagarth ndevos++
16:27 glusterbot hagarth: ndevos's karma is now 38
16:36 ira joined #gluster-dev
17:02 jobewan joined #gluster-dev
17:44 _Bryan_ joined #gluster-dev
17:59 lalatenduM joined #gluster-dev
18:16 jobewan joined #gluster-dev
19:20 lalatenduM kkeithley, ndevos http://review.gluster.org/#/c/8994/
19:22 jobewan joined #gluster-dev
19:47 lalatenduM joined #gluster-dev
21:20 jobewan joined #gluster-dev
21:23 badone joined #gluster-dev
21:37 badone joined #gluster-dev
21:37 davemc the etherpad is updated for tomorrows community meeting. Please feel free to review it and offer modifications
21:37 davemc Community meeting tomorrow at 12:00 UTC
21:37 davemc etherpad is at https://public.pad.fsfe.org/​p/gluster-community-meetings
22:05 badone joined #gluster-dev

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