Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2014-03-13

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

All times shown according to UTC.

Time Nick Message
00:04 kkeithley1 joined #gluster-dev
00:33 badone_ joined #gluster-dev
00:49 glusterbot` joined #gluster-dev
01:01 bala joined #gluster-dev
01:02 yinyin joined #gluster-dev
01:45 kdhananjay joined #gluster-dev
03:14 bharata-rao joined #gluster-dev
03:38 shubhendu joined #gluster-dev
03:47 hagarth joined #gluster-dev
04:01 mohankumar__ joined #gluster-dev
04:03 aravindavk joined #gluster-dev
04:11 itisravi joined #gluster-dev
04:40 ndarshan joined #gluster-dev
04:48 deepakcs joined #gluster-dev
04:53 spandit joined #gluster-dev
05:06 hagarth joined #gluster-dev
05:16 ppai joined #gluster-dev
05:17 vpshastry1 joined #gluster-dev
05:20 hagarth joined #gluster-dev
05:33 bala1 joined #gluster-dev
05:55 mohankumar__ joined #gluster-dev
05:57 lalatenduM joined #gluster-dev
06:05 raghu joined #gluster-dev
06:39 mohankumar__ joined #gluster-dev
06:48 mohankumar__ joined #gluster-dev
06:54 mohankumar__ joined #gluster-dev
07:12 edward1 joined #gluster-dev
07:21 pk1 joined #gluster-dev
07:33 chandan_kumar joined #gluster-dev
07:55 pk2 joined #gluster-dev
08:21 chandan_kumar lalatenduM, ping
08:22 chandan_kumar lalatenduM, i have installed gluster swift in a vm
08:22 chandan_kumar lalatenduM, i have added keystone config to /etc/swift/proxy-server.conf
08:23 chandan_kumar by using that document
08:23 chandan_kumar https://github.com/gluster/gluster-swift​/blob/master/doc/markdown/auth_guide.md
08:24 chandan_kumar when i am trying to list the containers: http://fpaste.org/84937/13946990/
08:25 chandan_kumar chandan_kumar, where to debug for this?
08:30 mohankumar__ joined #gluster-dev
08:35 lalatenduM chandan_kumar, I never used gluster-swift, ppai might know.
09:00 * jclift yawns
09:03 pk2 jclift: Soft kitty, Warm kitty, Little ball of fur. Happy kitty, Sleepy kitty, Purr Purr Purr
09:03 pk2 jclift: I hope you watch big bang theory
09:18 ndevos :)
09:18 pk2 ndevos: was that on my msgs?
09:20 ndevos pk2: yes
09:21 pk2 ndevos: I donno why jclift is not responding, hope he is sound asleep :-)
09:21 ndevos pk2: it's still early for jclift, 9:21 in London
09:22 pk2 ndevos: ah!
09:22 ndevos and I dont think jclift is a morning person ;)
09:23 pk2 ndevos: :-)
09:44 ajha joined #gluster-dev
10:05 jclift pk2: Actually, no I don't. :)
10:06 jclift ndevos: Yeah, I'm not a morning person normally.  Actually got up before noon today, so I'm here now for once.
10:06 jclift pk2: How can I help?
10:07 jclift Turns out there's a wikipedia article on it: http://en.wikipedia.org/wiki/Warm_Kitty
10:07 jclift No idea what it means tho. ;)
10:12 pk2 jclift: when you said you were yawning, this is what I remembered so I just wrote that for fun thats all. :-)
10:13 jclift :)
10:15 aravindavk joined #gluster-dev
10:22 social_ http://build.gluster.org/job/smoke/6891/console < this isn't my mistake, is it?
10:23 ndevos social_: no, unless you're playing with cmockery2?
10:24 ndevos using git submodules is really fragile :-/
10:26 ndevos social_: lets try again... http://build.gluster.org/job/smoke/6892/console
10:26 ndevos Running autogen.sh in cmockery2 ...
10:26 ndevos ./autogen.sh: line 117: ./autogen.sh: No such file or directory
10:27 ndevos that does not look promising :-/
10:30 social_ how is it done, you  get a patch add it to master and try to build?
10:30 jclift The way I fix that (manually) is to add "autogen.sh" and "m4" to the EXTRA_DIST line in cmockery/Makefile.am in local git (then run ./autogen.sh, etc)
10:30 jclift social_: ^^^
10:31 jclift The source of the bug is in cmockery2 git through, not GlusterFS.
10:32 ndevos jclift: but cmockery2 gets checked out, and the github repo has a autogen.sh file...
10:32 jclift ndevos: Sure.
10:32 jclift ndevos: the autogen.sh and m4 directories under cmockery2 are completely ignored on some platforms though if they're not explicitly called out
10:32 ndevos jclift: I've just done a 'sudo rm -rf /var/lib/jenkins/jobs/smoke/workspace/cmockery2' on build.gluster.org, and http://build.gluster.org/job/smoke/6893/console is progressing now
10:33 jclift ndevos: Interesting.  Cool. :)
10:34 ndevos jclift: did I already mention that I dont like git submodules because it breaks easily?
10:34 * jclift isn't a fan either
10:34 pk2 left #gluster-dev
10:35 pk2 joined #gluster-dev
10:35 jclift ndevos: I think your idea of cmocka-devel sounds like the way to go, as long as cmocka is compatible.
10:36 jclift lpabon mentioned cmocka doesn't have Jenkins integration yet, but did say it could be done
10:39 ndevos social_: it's fixed now, I've scheduled a regression test so that the failure from the smoke test gets dropped
10:39 ndevos you can watch http://build.gluster.org/j​ob/regression/3770/console in case you're interested :)
10:40 ndevos ah! it again fails to find autogen.sh :-/
10:41 jclift Hmmm, as a dodgy workaround, we could try moving the git submodule checkout to before the auto* running into autogen.sh, + add the cmockey2/autogen.sh + cmockery2/m4 dirs to the EXTRA_DIST line using sed.
10:41 jclift s/running into/running in/
10:42 ndevos jclift: or add a 'rm -rf cmockery2' in autogen.sh :)
10:42 jclift Heh, go for it. :)
10:43 jclift ndevos: We could back out the cmockery2 change in git too, until the problems are resolved.
10:43 ndevos jclift: its not as easy as that... all patches would need to be rebased to include that updated autogen.sh :-/
10:44 jclift *sigh*
10:45 jclift There don't seem to be any good choices here. :(
10:46 jclift Wish I'd tested the cmockery2 change on EL5 before verifying it.
10:46 jclift Would have seen this error then, and we'd be in an ok state. :/
10:46 ndevos I wish I'd seen the patch before it got merged :-/
10:53 ppai joined #gluster-dev
10:54 ndevos well, still does not work, it seems that autogen.sh is just sometimes not checked out or something, weirdness
10:54 spandit joined #gluster-dev
11:18 ajha joined #gluster-dev
11:22 social_ O.o
11:30 lalatenduM pk2, it awesome that u remember the  warm kitty song :)
11:30 lalatenduM I mean completely  :)
11:30 pk2 lalatenduM: googled :-P
11:30 lalatenduM pk2, LOL , I was impressed
11:37 hagarth joined #gluster-dev
11:53 mohankumar__ joined #gluster-dev
11:58 ppai joined #gluster-dev
12:04 ndevos kkeithley: I think I hit bug 983431 on my home 3.4 installation, should I add http://review.gluster.org/5319 to the backport wishlist? or was there a blocked bug?
12:04 glusterbot Bug https://bugzilla.redhat.com​:443/show_bug.cgi?id=983431 high, unspecified, ---, nsathyan, MODIFIED , DHT: NFS process crashed on a node in a cluster when another storage node in the cluster went offline
12:04 ndevos *blocker bug
12:06 shyam joined #gluster-dev
12:09 itisravi joined #gluster-dev
12:18 mohankumar__ joined #gluster-dev
12:21 social_ ndevos: still failed :) anyway how do you trigger rebuild? you have account in jenkins?
12:22 yinyin joined #gluster-dev
13:10 ndevos social_: yes, I've got an account so that I can help with troubleshooting tests etc...
13:17 pk1 joined #gluster-dev
13:20 lpabon joined #gluster-dev
13:26 kdhananjay joined #gluster-dev
13:30 social_ ndevos: ook I thought that jenkins account is linked with gerrit one but that's probably quite bad idea as that would mean arbitrary remote code execution for anyone who creates account
13:31 pk1 left #gluster-dev
13:31 ndevos social_: yeah, starting regression tests should be done with some care - not sure who manages the accounts and where you can apply for one
13:34 ndevos social_: just tried it again, removed cmockery2 in the jenkins workspace, but http://build.gluster.org/j​ob/regression/3772/console fails too
13:35 ndevos ping lpabon: there seems to be a weird issue in checking out and running ./autogen.sh in the cmockery2 repo :-/
13:35 ndevos lpabon: it causes errors in the smoke tests, and regression like http://build.gluster.org/j​ob/regression/3772/console
13:39 social_ ndevos: NP but I'd prefer review on that issue, it's again questionable solution imho
13:45 jclift ndevos: That's a weird error with the cmockery2 failure.
13:46 ndevos jclift: yes, and I have no idea how to get rid of that
13:46 jclift ndevos: It looks very similar to the "missing files in the tarball" problem I hit the other day (eg autogen.sh, m4 dir)
13:47 ndevos jclift: okay, but I do not think there is a 'make dist' in there...
13:47 jclift ndevos: Are you able to somehow break the build process mid step, or are the dirs from the broken build still on build.g.o?
13:48 jclift ndevos: eg check if there file is indeed missing from cmockery2 for some reason
13:50 ndevos social_: fwiw, I've added a note to that dht patch, but I'm not very familiar with dht's internals
13:52 lpabon ndevos: hi, yeah im investigating  the patch atm
13:52 social_ aww you are right I'm leaking dict_ref thre
13:52 ndevos lpabon: thanks for looking into it!
13:55 lpabon ndevos: social_: it looks to me like the fix jclift put in cmockery2 .. the fix is in the fix i have 7215 (but that one seems to be having some git issues, which I am checking out)
13:55 jclift ndevos: I'm tempted to put a patch up for review that adds basic python util functions in the tests/ subdirectory.  Python versions of the (presently bash based) cleanup functions and similar, so people can write their .t tests in Python.  And we people can debug them properly if they hit issues.
13:56 jclift ndevos: I've already tested running python based tests locally.  Works fine.  Prove doesn't know the difference. ;)
13:57 jclift lpabon: Good luck. :)
13:58 ndk joined #gluster-dev
13:58 lpabon :)
14:01 lpabon cross your fingers to make regression 3773 pass .. http://build.gluster.org/job/regression/3773/
14:02 ndevos jclift: I dont have a strong opinion on writing test scripts in python, shell works fine for me, but maybe others prefer python
14:02 lpabon <cough>python!<cough>
14:04 lpabon actually in gluster-swift we use a shell script to setup the functional test environment :-)
14:08 ndevos lpabon: would you move to cmocka when the changes from cmockery2 have been included?
14:09 lpabon there is a lot of enhancements in cmockery2 that would need to be moved to cmocka first
14:09 lpabon but sure, once all the enhancements are there, there is no issue (other than header file name changes and such).  But then we will need also to have all develoers install cmocka on their system
14:09 ndevos hmm, okay, but I think teh cmocka devs (Anderas Scheider?) are pretty open for including patches
14:10 lpabon I would delegate that answer to gluster devs
14:10 jclift lpabon: Having all the dev's install cmocka on their system is a completely minor thing.
14:10 jclift lpabon: Adding to developer dependencies isn't that much of a problem, as long as the dependency is widely available and easily consumable.
14:11 jclift (eg people already install stuff by cut-n-pasting from "how to do XYZ" web page)
14:11 * ndevos does not have a preference, but would like to drop as many bundled things as possible
14:11 lpabon ok
14:11 jclift lpabon: Adding to run time dependencies, for users, can be a pain though.
14:11 lpabon ndevos: i agree
14:12 jclift cmocka doesn't add any run time dependency does it?  eg this would have no effect on end user installation requirements?
14:12 ndevos lpabon: have you looked at the difficulty of porting the cmockery2 changes to cmocka already?
14:12 lpabon like i said in the patch comments, I'm planning to do both .. add cmockery2 as an rpm and send a patch to cmocka... let glusterfs then decide
14:13 lpabon ndevos: its not easy but it is not hard. probably around 3-5 days work
14:13 lpabon from my point of view adding cmockery2 as an rpm is a lot easier but there is always a delay from fedora to add it.
14:13 ndevos lpabon: okay, that sounds doable
14:13 lpabon Also?  is cmocka in RHEL?
14:14 ndevos no, cmocka would need to be put in epel, but that should be quick (1-2 days)
14:14 lpabon ok, sounds like we may be adding more work
14:15 lpabon ok, how about this... in the short term we stay w/ it as is,  and add the work to both cmocka and cmockery2 RPMs in fedora and RHEL/Centos
14:15 lpabon Let glusterfs devs then pick what they want to use
14:16 ndevos lpabon: pro/cons for any of those?
14:16 wushudoin joined #gluster-dev
14:16 ndevos lpabon: if you start a package review for cmockery2 now, I can do a review later today or tomorrow
14:16 lpabon For including the code today (what my current patch is doing) -- Pro: easy install for all OSs, easy development - Con: Its part of the repo
14:17 ndevos yes, including cmockery2 is fine, for now :)
14:17 tdasilva joined #gluster-dev
14:17 lpabon ndevos: For fedora?
14:17 ndevos lpabon: yes
14:17 lpabon ndevos: don't tempt me, :-)
14:17 jclift lpabon: Was kind of thinking you might want to get your cmockery2 improvements into cmocka, so you don't have to spend time on cmockery2 any more and can put time into other things.
14:18 jclift lpabon: Doesn't sound like the thought appeals to you though?
14:18 lpabon i think both do actually
14:18 jclift Heh. ;)
14:18 lpabon we can work on it together
14:18 lpabon ndevos: watch for my cmockery2 spec file tonighte :-)
14:19 lpabon ..actually by your tomorrow morning
14:19 ndevos lpabon: when you post a review request, just add me on CC and I should see it
14:19 lpabon awesome... your redhat mail or the other one used by gerrit?
14:20 ndevos lpabon: the redhat one, also for gerrit you should pick that one :)
14:20 lpabon weird thing is that it was not letting me.. there were three niels with redhat...i tried each one.. gave me an error on each
14:20 ndevos lpabon: you're a package maintainer already, right?
14:21 lpabon only the nonredhat one worked
14:21 lpabon yes--sh
14:21 lpabon i'm working (since sept 2013) to add gluster-swift as an rpm to fedora .. but that one has issues because of its dep on openstack swift
14:21 ndevos whats your fedora username?
14:21 lpabon lpabon
14:22 lpabon i don't think I have a lpabon @ fedora mail if that is what you are asking
14:22 lpabon i don't have my first repo yet either
14:23 lpabon but i can log into the fedora website to add my ssh key and such
14:23 ndevos hmpf, you're not in the packager group yet :-/
14:23 lpabon i don't think so.. i'm still a newbie at fedora packages :--(
14:23 ndevos https://admin.fedoraproject.o​rg/accounts/user/view/lpabon - and change lpabon to devos to see my groups
14:25 * ndevos looks for some magick powers
14:25 jobewan joined #gluster-dev
14:26 lpabon ok I see your groups
14:30 ndevos lpabon: can you refresh the page with your groups?
14:30 lpabon sweet!
14:31 lpabon now it shows Git commit group , bugzilla , contributor agreement, and cla
14:32 ndevos lpabon: yes, that means you do not need the sponsor for becoming a package maintainer anymore - I've got you sponsored, but don't mess up!
14:33 lpabon with great power comes great.. ah you know
14:37 ndevos lpabon: your regression test fails again, building the rpms in mock seems to be an issue?
14:37 lpabon yeah dumb issue
14:37 lpabon i have less files, so the %exclude is not finding anything to exclude :)
14:37 lpabon fixing now
14:37 ndevos ah, right
14:48 glusterbot` joined #gluster-dev
15:00 lpabon_ joined #gluster-dev
15:04 lpabon_ joined #gluster-dev
15:27 kkeithley1 joined #gluster-dev
15:30 deepakcs joined #gluster-dev
16:02 kkeithley_ did something happen to build.gluster.org (gerrit)?
16:07 kkeithley_ ^^^  anyone???
16:08 jclift ndevos: You were doing stuff with it?
16:09 jclift Hmmm, website isn't responding.  Guess that's what you're meaning?
16:09 jclift kkeithley_: ^
16:09 ndevos build.gluster.org works fine? review.gluster.org seems to be down?
16:09 kkeithley_ sorry, not build, review.gluster.org
16:11 ndk joined #gluster-dev
16:18 ira_ joined #gluster-dev
16:21 hagarth joined #gluster-dev
16:22 jclift hagarth: ping
16:22 jclift hagarth: review.gluster.org is down.  Do we have to wait for avati?
16:23 ndevos jclift: maybe its just an other patch-bomb?
16:31 hagarth jclift: checking
16:34 jclift hagarth: x
16:34 jclift tx
16:35 hagarth jclift: looks like we'll need Avati for accessing the hypervisor which hosts review.gluster.org
16:35 jclift k
16:36 * jclift idly wonders if it would be a good idea to setup review.gluster.org in rackspace or something, which would probably make for an admin console with access for multiple people
16:37 hagarth jclift: yeah, need to setup some time with johnmark to explore the rackspace route
16:37 jclift hagarth: He's passed the Rackspace credentials to purpleidea, so purpleidea might have some hands on idea of what's possible with the vm's there
16:38 hagarth jclift: ok, cool.
16:42 ndevos jclift: you like interesting bugs right? https://bugzilla.redhat.co​m/show_bug.cgi?id=1074023 might be something for you then
16:42 glusterbot Bug 1074023: high, unspecified, ---, ndevos, ASSIGNED , list dir with more than N files results in Input/output error
16:44 jclift Hmmm, BZ is very slow.
16:47 jclift ndevos: That's definitely a weird looking error.  Seems like you've already solved it though. :)
16:48 ndevos jclift: it's solved because I hit it at home too :D
16:48 jclift :)
17:23 jdarcy joined #gluster-dev
17:24 jdarcy Does anyone know what's going on with review.gluster.org?
17:25 ndevos no, it seems we're waiting for avati to look at the hypervisor :-/
17:28 jdarcy ndevos: Thanks.  BTW, I finally sent your shirt yesterday.  Sorry for the delay.
17:30 ndevos jdarcy: oh, wow, thanks! A delay is no problem at all :)
17:51 lpabon joined #gluster-dev
18:06 lalatenduM joined #gluster-dev
18:20 lpabon I'm trying to see if we can run the regression on our development systems.  The jenkins job 'regression' uses two scripts which are not on the repo.  Can we move /opt/qa/build.sh and /opt/qa/regression.sh into the glusterfs repo?
18:24 johnmark ndevos: ping
18:24 johnmark ndevos: you there?
18:29 johnmark jclift: yes, need to set it up in rax
18:29 johnmark hagarth: I'll send you creds
18:35 johnmark well, seems to be working now..
18:36 lpabon finally
18:44 jdarcy joined #gluster-dev
18:56 kkeithley_ who fixed it?
19:55 ndevos johnmark: I'm here now?
20:18 jclift Yep, it's online now.
20:33 jclift lpabon: With that new patch of yours, it doesn't include the cmockery/m4 directory, but your Makefile.am still has "ACLOCAL_AMFLAGS = -I m4".
20:33 jclift lpabon: Have you tested the patch on EL5, just to be sure?
20:33 lpabon libtoolize creates files in the m4 directory
20:34 * jclift is kinda paranoid after last time
20:34 jclift k
20:34 lpabon no, i should do that.. you are correct to be paranoid.. i'll try in  5 mins
20:34 jclift :)
20:38 badone_ joined #gluster-dev
20:49 lpabon jclift: I built glusterfs with my latest patch..
20:49 lpabon did a ./autogen.sh && ./configure --enable-fusermount && make -j 8 check
20:49 lpabon all worked
20:51 kkeithley_ purpleidea: ping
21:00 jclift lpabon: Try making rpms with it, to be sure
21:00 jclift eg make dist, and make glusterrpms
21:00 lpabon ok, I will try now
21:00 jclift Tx.  It's the rpm making that was barfing last time
21:00 kkeithley_ Nuke it from orbit, it's the only way to be sure.
21:00 lpabon where do i get git for centos 5?
21:01 jclift EPEL
21:01 kkeithley_ epel?
21:01 lpabon k
21:01 jclift Hmmm
21:01 jclift I know I had this problem too.  Pretty sure it was EPEL
21:01 * jclift checks
21:01 jclift http://www.gluster.org/community/do​cumentation/index.php/CompilingRPMS
21:01 jclift Yeah
21:02 jclift curl -OL http://download.fedoraproject.org/pub/ep​el/5/x86_64/epel-release-5-4.noarch.rpm
21:02 jclift sudo yum -y install epel-release-5-4.noarch.rpm --nogpgcheck
21:02 jclift lpabon: Might just be easier to cut-n-paste your way along the page ^
21:21 lpabon jclift: passed. all rpms were created
21:21 lpabon jclift: example:  Wrote: /root/glusterfs/extras/LinuxRPM/rpmbuild/RPMS/​x86_64/glusterfs-3git-1.el5.centos.x86_64.rpm
21:45 jclift lpabon: Excellent. :)
22:12 kkeithley1 joined #gluster-dev

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