Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2014-05-28

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

All times shown according to UTC.

Time Nick Message
00:36 awheeler joined #gluster-dev
00:42 yinyin_ joined #gluster-dev
00:48 yinyin- joined #gluster-dev
01:05 badone joined #gluster-dev
01:14 bala joined #gluster-dev
01:20 pranithk joined #gluster-dev
01:20 pranithk JustinClift: ping
01:56 jobewan joined #gluster-dev
02:00 pranithk left #gluster-dev
02:51 bharata-rao joined #gluster-dev
03:37 itisravi joined #gluster-dev
03:46 nishanth joined #gluster-dev
03:47 shubhendu joined #gluster-dev
03:51 kanagaraj joined #gluster-dev
03:54 ppai joined #gluster-dev
03:54 shubhendu_ joined #gluster-dev
04:36 ndarshan joined #gluster-dev
04:43 krishnan_p joined #gluster-dev
04:43 hagarth joined #gluster-dev
04:53 spandit joined #gluster-dev
05:00 kdhananjay joined #gluster-dev
05:01 bala joined #gluster-dev
05:06 kshlm joined #gluster-dev
05:12 aravindavk joined #gluster-dev
05:24 skoduri joined #gluster-dev
05:27 lalatenduM joined #gluster-dev
05:40 baojg joined #gluster-dev
05:45 vpshastr1 joined #gluster-dev
05:47 raghu joined #gluster-dev
05:47 baojg_ joined #gluster-dev
05:50 baojg_ joined #gluster-dev
05:54 krishnan_p joined #gluster-dev
05:55 hagarth joined #gluster-dev
05:59 awheeler joined #gluster-dev
06:31 hagarth joined #gluster-dev
06:34 vpshastr1 joined #gluster-dev
06:43 krishnan_p joined #gluster-dev
06:46 kdhananjay joined #gluster-dev
07:02 vpshastr1 joined #gluster-dev
07:10 kaushal_ joined #gluster-dev
07:12 vpshastry2 joined #gluster-dev
07:42 kdhananjay joined #gluster-dev
07:46 kshlm joined #gluster-dev
07:51 baojg joined #gluster-dev
07:56 vpshastry joined #gluster-dev
07:58 awheeler joined #gluster-dev
08:21 rgustafs joined #gluster-dev
08:52 nishanth joined #gluster-dev
09:00 scuttle_ joined #gluster-dev
09:19 vpshastry joined #gluster-dev
09:22 nishanth joined #gluster-dev
09:35 [o__o] joined #gluster-dev
09:42 vpshastry joined #gluster-dev
09:47 vpshastry joined #gluster-dev
10:32 krishnan_p joined #gluster-dev
10:37 scuttle_ joined #gluster-dev
11:00 tdasilva joined #gluster-dev
11:02 ppai joined #gluster-dev
11:10 baojg joined #gluster-dev
11:19 [o__o] joined #gluster-dev
11:22 hagarth joined #gluster-dev
11:42 ira joined #gluster-dev
11:54 baojg joined #gluster-dev
12:02 itisravi joined #gluster-dev
12:28 edward1 joined #gluster-dev
12:31 tdasilva joined #gluster-dev
12:44 tdasilva left #gluster-dev
12:46 tdasilva joined #gluster-dev
12:59 awheeler joined #gluster-dev
12:59 [o__o] joined #gluster-dev
13:25 [o__o] joined #gluster-dev
13:51 hagarth joined #gluster-dev
14:01 itisravi joined #gluster-dev
14:15 kshlm joined #gluster-dev
14:21 wushudoin joined #gluster-dev
14:26 pranithk joined #gluster-dev
14:26 pranithk JustinClift: ping
14:26 JustinClift pranithk: pong
14:26 pranithk JustinClift: Did you get a chance to run debug build?
14:26 pranithk JustinClift: for regressions I mean
14:27 JustinClift Hmmmm, I forgot to put those options into the build script.
14:27 JustinClift However, I can do it shortly.  Just need to finish two emails first. :)
14:27 JustinClift I've only got one bug blocking the regressions-on-rackspace.
14:27 pranithk JustinClift: oh. I would love to see how they turn out :-). Sure, send a mail on gluster-devel with the results. It should be interesting :-)
14:28 pranithk JustinClift: oh!
14:31 JustinClift pranithk: This is the only one left: http://fpaste.org/105355/01287487/
14:31 glusterbot Title: #105355 Fedora Project Pastebin (at fpaste.org)
14:32 JustinClift pranithk: It happens with both of the jenkins slaves I have setup, so I think it's something to do with the configuration of the VM's ip addresses, rather than a race condition thing.
14:33 JustinClift pranithk: kshlm mentioned he'd be able to take a look.  I'm hoping for today. :)
14:34 pranithk JustinClift: Cool!
14:38 kshlm JustinClift, I'm gonna take a look now.
14:38 pranithk JustinClift: On my fedora vm 15 tests fail consistently..
14:38 JustinClift pranithk: Interesting.  That's with the mempool disabled, and debug build turned on?
14:38 pranithk JustinClift: Even in the mail Raghavendra Talur sent about running regressions using docker... some of the tests are failing
14:38 pranithk JustinClift: Nope. Normal Rpms
14:39 JustinClift pranithk: Is rpm.t in your consistent failure set?
14:39 * JustinClift has found a few configuration things that need to be "just right" for that one to work
14:40 pranithk pranithk: http://paste.fedoraproject.org/105362/12879941
14:40 glusterbot Title: #105362 Fedora Project Pastebin (at paste.fedoraproject.org)
14:40 JustinClift Interesting
14:40 JustinClift k, that's a better list than mine.
14:41 JustinClift I haven't done a large scale test in a while.  Been either a) distracted by non-test stuff, or b) trying to get the test infrastructure working
14:41 pranithk JustinClift: I am working on them one after another when I get some time... It may take a long while before all of these are fixed properly...
14:41 JustinClift Heh.  Painful isn't it? :/
14:42 pranithk JustinClift: I actually enjoy debugging these... Most of them would be some kind of races.. :-)
14:42 JustinClift pranithk: ;)
14:42 JustinClift kshlm: If the solution turns out to just be "we need to configure the VM's a bit differently" that's probably acceptable
14:45 kshlm Ok. Is it okay if I mess with the system?
14:45 kshlm I want to manually run the test.
14:45 JustinClift kshlm: Yeah, go for it
14:45 kshlm Cool.
14:45 JustinClift I have a few other systems ready for me to setup :)
14:46 scuttle_ joined #gluster-dev
14:59 ndevos JustinClift: meeting time?
15:00 JustinClift *** Gluster Community Meeting time is NOW: #gluster-meeting on irc.freenode.net ***
15:00 JustinClift ndevos: Yep :)
15:08 kshlm JustinClift, looks like the gnfs authentication is not handling the hostname of the vm correctly. I'm digging into it now.
15:09 JustinClift kshlm: Thanks. :)
15:13 baojg joined #gluster-dev
15:21 kshlm JustinClift, Found the problem. NFS authentication doesn't handle IPV6 addresses!!
15:23 kshlm If you can disable ipv6 on the vms, we are ready to go. But we do need to fix the authentication.
15:24 JustinClift kshlm: Ahhh, yeah.  IPv6 can be disabled completely on them.  I'll do that. :)
15:32 kshlm JustinClift, I'm done with the vm.
15:32 JustinClift kshlm: Thanks heaps for looking into this.  Makes a difference. :)
15:32 kshlm I'm gonna raise a bug for this issue.
15:33 skoduri joined #gluster-dev
15:41 kshlm JustinClift, I think the problem is bigger. Just disabling IPV6 may not work.
15:42 kshlm The VM had 2 networks attached, this could lead to failures for that particular test as well.
15:42 kshlm I checked the authentication code and this is a possiblity for failure as well.
15:43 kshlm Could these VMs have just 1 IPv4 only network configured?
15:43 * kshlm is sorry for more bad news.
15:44 JustinClift kshlm: I don't know.  Did you try disabling the network that's not the public interface?
15:44 JustinClift kshlm: A lot of hosting companies create 2 interfaces on their VM's
15:44 JustinClift Interface 1) Public one, for things to access it
15:45 JustinClift Interface 2) Private internal-to-hosting-company one, so several VM's owned by same customer can communicate without being charged for the bandwidth
15:45 JustinClift (they only charge for bandwidth on the public interface)
15:45 JustinClift kshlm: Since our VM's only need the 1 interface, I think we'll be ok
15:47 kshlm Hmm, but if we shut down the internal interface, we'll be using the public interface during the tests.
15:47 JustinClift Sure
15:47 JustinClift Traffic doesn't actually get sent out over it though does it?
15:47 JustinClift eg nothing actually leaves the VM using it
15:47 kshlm Hopefully not.
15:48 JustinClift ;)
15:48 JustinClift We'll be fine.  I think. ;)
15:49 kshlm Ok. I'll test it, after I get some dinner.
15:49 JustinClift :)
15:49 kshlm Could you disable IPv6 on the vm in the meantime?
15:49 JustinClift Sure
15:49 JustinClift I'll need to reboot it though
15:50 JustinClift Gimme 5 mins
15:50 kshlm Not a problem. I'll be away long enough for it to come back up.
15:50 JustinClift Need to send out the meeting summary email :)
15:50 JustinClift np
16:32 JustinClift kshlm: IPv6 disabled, and the other interface is gone.
16:33 JustinClift kshlm: Looks good now: http://fpaste.org/105406/14012948/
16:33 glusterbot Title: #105406 Fedora Project Pastebin (at fpaste.org)
16:34 JustinClift kshlm: I'm going to kick off a full regression run on it, and see how it goes.
16:35 kshlm That's good news.
16:36 kshlm I'll raise a bug with the problem anyway. We need better support ipv6.
16:36 JustinClift Sounds good. :)
16:36 kshlm s/support/support for/
16:36 JustinClift Sure
16:41 ndevos kshlm: that is actually one of the proposed features for 3.6 :)
16:42 JustinClift :)
16:42 kshlm I'm afraid we aren't in a really good state w.r.t ipv6.
16:42 kshlm I don't believe any of the developer ever uses ipv6 during developement.
16:43 kshlm I tried once, but gave up because I couldn't setup an environment.
16:43 JustinClift I guess we'll need someone who knows IPv6 to really get stuck into it + make it basically work, then document how other devs should setup their system to use it
16:44 vpshastry1 joined #gluster-dev
16:46 kshlm We need a volunteer. I would do it, but I cannot setup an IPv6 environment.
16:46 kshlm I could try if I got access to an IPv6 environment.
16:49 JustinClift kshlm: Rackspace definitely has IPv6 available to all of its VMs
16:49 JustinClift kshlm: You should also be able to setup IPv6 environment in VM's on your desktop/laptop/etc
16:50 * JustinClift doesn't know if we use IPv6 in RH internal network
16:50 kshlm I tried that once and failed. I screwed up the existing network on my laptop badly and stopped after that.
16:50 JustinClift ;)
16:51 kshlm Maybe we should run the regression tests on a Rackspace VM with only IPv6 available. It'll let us know where we stand.
16:52 JustinClift kshlm: We can do that
16:52 * kshlm expects a lot of tests to fail.
16:52 JustinClift We could set up an IPv6 slave pretty easily
16:52 * JustinClift isn't sure how to SSH to an IPv6 slave now I think about it...
16:52 JustinClift Hmmm, is Jenkins IPv6 capable?
16:53 kshlm Once we make that public, we'll hopefully have more interest in fixing those problems.
16:53 JustinClift Good idea
16:53 kshlm I have no clue.
16:59 hchiramm_ joined #gluster-dev
16:59 lpabon joined #gluster-dev
17:00 JustinClift lpabon: Good news.  It's possibly to run simultaneous regression jobs on different hosts
17:00 lpabon JustinClift: really!? awesome!
17:00 JustinClift lpabon: eg: http://build.gluster.org/job/jc-regression1
17:00 glusterbot Title: jc-regression1 [Jenkins] (at build.gluster.org)
17:00 JustinClift Yeah.  The parallel executors thing is working fine.
17:01 JustinClift (configuring the slaves so GlusterFS works on them has taken some effort, but seems almost ok now)
17:01 lpabon nice! now if they only passed ;-)
17:01 JustinClift ;)
17:04 kshlm JustinClift, FYI https://bugzilla.redhat.com/show_bug.cgi?id=1102293 , this is the bug regarding nfs auth issues.
17:04 glusterbot Bug 1102293: unspecified, unspecified, ---, rwheeler, NEW , NFS subdir authentication doesn't correctly handle multi-(homed,protocol,etc) network addresses
17:08 lkoranda joined #gluster-dev
17:33 hchiramm_ joined #gluster-dev
18:01 lpabon joined #gluster-dev
18:04 lpabon joined #gluster-dev
18:06 [o__o] joined #gluster-dev
18:08 JustinClift kshlm: Wow.  Full regression run passed: http://build.gluster.org/job/jc-regression1/15/console
18:08 glusterbot Title: jc-regression1 #15 Console [Jenkins] (at build.gluster.org)
18:09 JustinClift I've been using Humble's CR 7881 as a test one, since that's no code change.
18:36 JustinClift Wow.  Two passing regressions, one on each slave.
18:36 JustinClift Lets try for a hat trick...
18:56 kkeithley are you feeling lucky?
18:57 JustinClift :)
19:02 tdasilva joined #gluster-dev
19:37 tdasilva joined #gluster-dev
19:56 awheeler joined #gluster-dev
20:52 JustinClift kkeithley: I've just chmod -R g+w on /usr/local/* no the Lab mac mini.  If it makes things go weird you can blame me. ;)
20:52 JustinClift s/no/on/
20:58 JustinClift kshlm: Btw: http://jenkins-ci.361315.n4.nabble.com/IPv6-Support-Compliance-td4640285.html
20:58 glusterbot Title: Jenkins users - IPv6 Support/Compliance (at jenkins-ci.361315.n4.nabble.com)
20:59 JustinClift kshlm: In theory, Jenkins should work with IPv6
21:02 yinyin_ joined #gluster-dev
21:04 tdasilva left #gluster-dev
22:13 yinyin joined #gluster-dev
22:57 awheeler joined #gluster-dev

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