Camelia, the Perl 6 bug

IRC log for #gluster-dev, 2013-09-20

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

All times shown according to UTC.

Time Nick Message
02:29 bharata-rao joined #gluster-dev
03:05 kshlm joined #gluster-dev
03:23 itisravi joined #gluster-dev
03:30 shubhendu joined #gluster-dev
04:20 bulde joined #gluster-dev
04:22 kanagaraj joined #gluster-dev
04:26 ppai joined #gluster-dev
04:38 lalatenduM joined #gluster-dev
04:40 ndarshan joined #gluster-dev
04:45 bala joined #gluster-dev
04:52 ndarshan joined #gluster-dev
05:00 lalatenduM joined #gluster-dev
05:08 hagarth joined #gluster-dev
05:10 ndarshan joined #gluster-dev
05:20 mohankumar joined #gluster-dev
05:30 aravindavk joined #gluster-dev
05:31 ajha joined #gluster-dev
05:55 raghu` joined #gluster-dev
06:09 rgustafs joined #gluster-dev
06:42 ababu joined #gluster-dev
07:02 an joined #gluster-dev
07:38 bharata-rao unable to start an existing volume after moving to glusterfs-3.4.1qa2 -- http://dpaste.com/1388630/
07:39 bharata-rao hagarth, ^
07:45 hagarth bharata-rao: do you notice any errors in the brick log file?
08:12 bulde joined #gluster-dev
08:16 badone_ joined #gluster-dev
08:30 bharata-rao hagarth, something to do with changelog xlator, brick log here: http://dpaste.com/1388667/
08:46 kshlm joined #gluster-dev
08:47 hagarth bharata-rao: were you using mainline before moving to 3.4.1qa2?
08:47 bharata-rao hagarth, yes
08:47 hagarth mainline has changelog xlator loaded by default in the volfile
08:48 hagarth 3.4.1qa2 doesn't have it
08:48 bharata-rao hagarth, oho
08:48 bharata-rao hagarth, guess I have to turn off changelog then
08:48 hagarth bharata-rao: you can peform gluster volume reset <volname> and restart the volume
08:48 bharata-rao hagarth, ok
08:49 bharata-rao hagarth, nice worked, thanks
09:23 an_ joined #gluster-dev
09:48 hagarth bharata-rao: how do you get gluster logs with qemu - libgfapi?
09:50 ndarshan joined #gluster-dev
09:51 bharata-rao hagarth, I  am using only error logs (GF_LOG_ERROR) and it comes on /dev/stderr
10:03 hagarth bharata-rao: I have a patch that creates a logfile for libgfapi if you pass NULL logfile in glfs_set_logging()
10:03 hagarth http://review.gluster.org/#/c/5910/1/api/src/glfs.h
10:04 hagarth bharata-rao: can we consider using this in qemu gluster driver? that way we'll be better equipped to troubleshoot problems coming from glusterfs.
10:06 bharata-rao hagarth, let me check
10:07 bharata-rao hagarth, right now qemu is using "-"
10:07 bharata-rao hagarth, which is /dev/stderr
10:08 bharata-rao hagarth, but yes it is better to have a standard log file
10:10 hagarth bharata-rao: once this gets merged, I can send out a patch for qemu.
10:10 bharata-rao hagarth, while we are at this, can this be also fixed so that we can use standard GF_LOG_* macros from qemu instead of hardcoded values
10:11 bharata-rao hagarth, https://bugzilla.redhat.com/show_bug.cgi?id=860205
10:11 glusterbot Bug 860205: medium, medium, ---, kaushal, ASSIGNED , libgfapi: Make GF_LOG_* macros available to the users of libgfapi
10:11 hagarth yeah .. that would be useful.. will get that done.
10:11 bharata-rao hagarth, if that's not too much of an effort, both the above additions to qemu can be sent in one patch
10:18 hagarth sure, that should not be a problem.
10:50 rgustafs joined #gluster-dev
11:21 vshankar joined #gluster-dev
11:43 ndarshan joined #gluster-dev
12:16 mohankumar during make install i get following error: py-compile: Missing argument to --destdir
12:16 mohankumar hagarth: bulde: any idea? ^^
12:27 hagarth mohankumar: is that a fatal error? I don't think so.
12:27 mohankumar hagarth: yes, make install fails
12:28 mohankumar so i could not run glusterd
12:33 vshankar joined #gluster-dev
12:37 hagarth mohankumar: I get it all the time .. but make install succeeds
12:39 ndarshan joined #gluster-dev
12:53 mohankumar hagarth: some how its not installing glusterd and other files
12:53 mohankumar so that machine is totally not usable
12:57 awheeler joined #gluster-dev
12:57 awheeler joined #gluster-dev
13:02 hagarth1 joined #gluster-dev
13:44 jclift joined #gluster-dev
14:20 lpabon joined #gluster-dev
14:28 ndk joined #gluster-dev
14:31 jclift ndevos: Thanks for that updated kernel.  Massive improvement.  Using touchpad at office now, no longer need a separate mouse/keyboard.
14:31 jclift ndevos: Updated the BZ to point people at the koji build, and also tweeted about it.  Hopefully other people find it useful too. :)
14:32 ndevos jclift: awesome! I just read your email, I'll respond later to that - first need to finish a call
14:35 jclift ndevos: No worries at all. :)
14:41 jclift ndevos: Heh, already getting other reports of success using that kernel on the BZ.
15:06 kkeithley jclift, ndevos: your changes for the test packaging make the mock build for el5 in rpm.t fall down
15:07 kkeithley Installed (but unpackaged) file(s) found:
15:08 kkeithley /glusterfs/run-tests.sh
15:08 kkeithley ...
15:08 kkeithley /glusterfs/tests/volume.rc
15:08 kkeithley Child return code was: 1
15:08 kkeithley EXCEPTION: Command failed. See logs for output.
15:08 kkeithley :-(
15:08 kkeithley see /d/build/mock/epel-5-x86_64/result/build.log on build.gluster.org
15:09 kkeithley I guess none of the regressions run since the change was committed on HEAD have actually run rpm.t
15:12 kkeithley i.e. didn't change any Makefile.{am,in}, etc.
15:18 ndevos kkeithley: I think bulde and y4m4 solved that?
15:19 kkeithley I rebased my tree two or three hours ago. Did they fix it since then? I'll try rebasing again.
15:20 ndevos kkeithley: I think they did that yesterday evening, on build.gluster.org - sorry I have no further details
15:25 * jclift points the finger at this change set at ndevos. Nyah Nyah. :)
15:25 jclift kkeithley: Actually, I just looked over it a bit, but didn't write any code for it.
15:26 * kkeithley thinks don't fix the blame, fix the problem
15:26 jclift Doh
15:26 kkeithley ;-)
15:27 kkeithley was there a spec change? (I haven't looked yet)
15:27 * kkeithley thinking maybe you know off the top of your head
15:27 jclift kkeithley: Yeah, there was an addition of the package: http://review.gluster.org/#/c/5674/11/glusterfs.spec.in
15:27 * kkeithley looks, and yes
15:31 * kkeithley wonders if all the Requires are available in el5
15:33 kkeithley e.g. perl(App:Prove)
15:34 * ndevos wonders where y4m4 is when you need to blame hime ^W^W^Whim to fix it
15:36 ndevos kkeithley: the issue is that %{_datarootdir} is not defined on RHEL5
15:36 kkeithley I think in my tree I'll just %if it out on rhel < 6.
15:36 kkeithley oh, okay
15:36 ndevos at least, thats is my intepretation of this error:
15:36 ndevos error: File must begin with "/": %{_datarootdir}/glusterfs/*
15:36 ndevos error: File must begin with "/": %{_datarootdir}/glusterfs/tests/basic/rpm.t
15:38 ndevos oh, btw, anyone reading this, my team is looking for highly technical support engineers (kernel internals, filesystem development, systemtap debugging, ... etc), ping/email me if you're interested :D
15:39 ndevos kkeithley: will you file a patch for that, or shall I?
15:40 kkeithley do you think %if-ing it out on rhel < 6 is the right fix?
15:41 kkeithley s/think/agree/
15:43 * kkeithley doesn't care who files the patch. I've got hours to go here and it's getting close to Beer:30 where you are. I can do it if you want
15:43 kkeithley once we agree on the fix
15:46 ndevos I think defining _datarootdir is the way to go
15:47 kkeithley okay
15:49 kkeithley you don't know what that should be, off the top of your head, do you?
15:49 ndevos rpm -E '%_dataorootdir' ?
15:49 * ndevos tries
15:49 kkeithley /usr/share
15:49 kkeithley ?
15:49 ndevos yeah
16:04 kkeithley okay, I'll try it in my tree that I'm working in right now and if it works I'll enter a separate patch for it. I'll tie it to the existing BZ for the regression test package.
16:05 kkeithley but now it's lunch time
16:06 hagarth joined #gluster-dev
16:08 ndevos sounds good to me, thanks and enjoy your lunch :)
16:39 lalatenduM joined #gluster-dev
16:40 jclift ndevos: With that "my team is looking for..." thing, tweet it.
16:40 jclift (as well) :D
16:41 jclift left #gluster-dev
16:41 ndevos hah, yeah, but I guess I need to digg up the link with the job description for that to make it more effective :)
16:56 awheeler joined #gluster-dev
17:02 awheeler_ joined #gluster-dev
17:06 kkeithley ndevos: it's not just the glusterfs.spec.in that needs _datarootdir, theres $(datarootdir) in the autoconf'd Makefile
17:07 kkeithley and automake is barfing on my attempts to add ifndef datarootdir\ndatarootdir = $(prefix)/share\nendif
17:07 kkeithley to the Makefile.am or the the tests.mk
17:09 bulde joined #gluster-dev
17:12 kkeithley do you really think it's that important to have the regression-test RPM for el5?
18:11 an joined #gluster-dev
23:51 ndevos kkeithley: I'm fine with disabling the package for now, I (or y4m4) can have a go at it later

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