Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2014-02-21

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

All times shown according to UTC.

Time Nick Message
00:34 shyam joined #gluster-dev
00:47 overclk joined #gluster-dev
00:49 johnmark_ joined #gluster-dev
00:51 jclift joined #gluster-dev
02:58 bharata-rao joined #gluster-dev
03:22 shubhendu joined #gluster-dev
03:36 itisravi joined #gluster-dev
03:49 kanagaraj joined #gluster-dev
03:59 shyam joined #gluster-dev
04:04 ndarshan joined #gluster-dev
04:09 jobewan joined #gluster-dev
04:23 kdhananjay joined #gluster-dev
04:26 jobewan joined #gluster-dev
04:39 ppai joined #gluster-dev
04:39 mohankumar__ joined #gluster-dev
04:45 spandit joined #gluster-dev
04:54 mohankumar__ joined #gluster-dev
04:54 hagarth joined #gluster-dev
05:00 bala joined #gluster-dev
05:01 jobewan joined #gluster-dev
05:13 mohankumar__ joined #gluster-dev
05:17 ajha joined #gluster-dev
05:19 mohankumar__ joined #gluster-dev
05:35 badone joined #gluster-dev
05:44 lalatenduM joined #gluster-dev
05:45 bharata-rao joined #gluster-dev
05:54 ajha joined #gluster-dev
06:00 aravindavk joined #gluster-dev
06:13 mohankumar__ joined #gluster-dev
06:14 badone joined #gluster-dev
06:25 mohankumar__ joined #gluster-dev
06:31 mohankumar__ joined #gluster-dev
06:45 bharata-rao joined #gluster-dev
07:04 mohankumar__ joined #gluster-dev
07:14 mohankumar__ joined #gluster-dev
07:20 aravindavk joined #gluster-dev
07:21 shubhendu joined #gluster-dev
07:21 ndarshan joined #gluster-dev
07:23 kanagaraj joined #gluster-dev
07:24 bharata-rao joined #gluster-dev
07:24 mohankumar__ joined #gluster-dev
07:30 bala joined #gluster-dev
07:33 mohankumar__ joined #gluster-dev
07:46 bala joined #gluster-dev
08:02 kanagaraj joined #gluster-dev
08:03 ndarshan joined #gluster-dev
08:12 mohankumar joined #gluster-dev
08:12 badone joined #gluster-dev
08:19 aravindavk joined #gluster-dev
08:20 ndevos jclift, kkeithley1: hmm, nightly build of master failed with 'configure: error: C compiler cannot create executables', and that includes f1f91f8 'build: Don't discard CFLAGS in configure'
08:24 mohankumar joined #gluster-dev
08:45 shubhendu joined #gluster-dev
09:00 mohankumar joined #gluster-dev
09:12 ndarshan joined #gluster-dev
09:16 ajha joined #gluster-dev
09:17 bala joined #gluster-dev
09:18 mohankumar joined #gluster-dev
09:29 kanagaraj joined #gluster-dev
09:33 mohankumar__ joined #gluster-dev
09:41 bharata_ joined #gluster-dev
09:43 ndarshan joined #gluster-dev
09:48 bala joined #gluster-dev
09:52 an_ joined #gluster-dev
10:24 mohankumar__ joined #gluster-dev
11:05 kanagaraj joined #gluster-dev
11:13 raghu joined #gluster-dev
11:16 badone joined #gluster-dev
11:47 an_ joined #gluster-dev
11:55 itisravi joined #gluster-dev
12:07 edward2 joined #gluster-dev
12:25 mohankumar joined #gluster-dev
12:34 mohankumar joined #gluster-dev
12:37 jclift ndevos: Damn. :(
12:39 jclift Found a small gotcha with the test suite last night.  My home dir didn't have world readable permissions, so some of the tests that use mock (eg rpm.t) were failing.
12:45 ira joined #gluster-dev
12:58 * jclift will add something about that to the "Using the Gluster Test Framework page"
13:05 mohankumar joined #gluster-dev
13:07 jclift Making progress: http://www.gluster.org/community/documentatio​n/index.php/Using_the_Gluster_Test_Framework
13:08 jclift I'll need to convert that to asciidoc and put it on the new site page when it's a bit mre fleshed out.
13:19 hagarth joined #gluster-dev
13:38 mohankumar joined #gluster-dev
15:28 jclift On the "good news" side, my glupy change doesn't need any adjustments to the glupy.t test.  It runs fine.
15:44 an_ joined #gluster-dev
15:55 ndevos jclift: you have RDMA hardware around, yes? Do you know if a current glusterfs version works over rdma? (not IPoIB)
15:57 an_ joined #gluster-dev
15:59 jobewan joined #gluster-dev
16:50 bala joined #gluster-dev
17:23 jclift ndevos: Yeah, I have two low end boxes here with RDMA gear in it.  Haven't fired them up at all this year though, and I haven't really touched RDMA stuff in Gluster in ages.
17:23 jclift ndevos: It's last known state was "needs work", so I wouldn't advise running it in production.
17:23 jclift ndevos: Perf issues, and weird crashes.
17:25 ndevos jclift: okay, any idea if there are detailed bug reports? I currently only have bug 878883
17:25 glusterbot Bug https://bugzilla.redhat.com​:443/show_bug.cgi?id=878883 unspecified, medium, ---, rgowdapp, ASSIGNED , Fuse mount hangs for a volume with RDMA transport
17:26 ndevos jclift: I'd like to know if upstream glusterfs is doing better than the current rhs version...
17:28 jclift ndevos: It's unlikely to be doing better than the RHS version.  No-one's really been looking at the RDMA side of things in ages.
17:28 jclift ndevos: Pretty sure there's a BZ around about NFS problems with RDMA too. I'll take a look, I should be able to find it.
17:28 ndevos jclift: is that your next project?
17:29 ndevos jclift: that would be awesome! any comments in that bz are appreciated :)
17:29 jclift ndevos: It's on my list of things-i'd-like-to-get-working in gluster
17:30 ndevos jclift: you're almost finished with Glupy now, and you will need to test GlusterFlow with RDMA to see if it can keep up with the load ;)
17:30 jclift :p
17:30 jclift But yeah, you're probably right.
17:31 jclift ndevos: This is the one I was thinking of: https://bugzilla.redhat.com/show_bug.cgi?id=978205
17:31 glusterbot Bug 978205: medium, unspecified, ---, vagarwal, ASSIGNED , NFS mount failing for several volumes with 3.4.0 beta3.  Only last one created can be mounted with NFS.
17:32 jclift ndevos: If you need IB enabled hardware, there's some in beaker that I think isn't being used atm.
17:32 jclift ndevos: Can dig up the details if it's useful
17:32 ndevos jclift: hmm, yes, that might be related
17:33 jclift ndevos: Upstream actively discourages use of RDMA, to the point we put a warning into the cli when someone tries to create an RDMA volume
17:35 ndevos jclift: okay, we normally advise people to use IPoIB and be done with it, but still some people would like to see RDMA support working
17:35 jclift Yeah.  If we can sort out the problems it's not a bad idea.
17:36 ndevos I've never done anything with IB, so maybe its something I should try to have a look at rather sooner than later
17:37 jclift ndevos: Emailed the host names to you.
17:37 jclift kkeithley1 would know if anyone's using them, etc.
17:37 jclift Fck
17:37 jclift Regression testing failed on my glupy patch again.
17:38 * jclift sighs
17:38 jclift I have no idea how to debug, as it works fine here on CentOS 6.5.  I've tested it, and it's a damn simple test. :(
17:39 jclift Ahhh, it's probably just the regression host being weird again.  It's doing the 'C compiler cannot create execytables' thing again on rpc-coverage.t.
17:41 jclift hagarth: Any thoughts about the RPM changes discussed in the meeting the other day?
17:42 jclift eg Moving Glupy into a glusterfs-extra-xlators rpm?
17:43 ndevos jclift: glusterfs-extra-xlators sounded as "accepted" to me :)
17:44 ndevos jclift: fwiw, I'm hitting the 'C compiler cannot create execytables' in the nightly builds on build.gluster.org too... I have not investigated yet why that would be the case (its not in rpm.t/mock)
17:44 jclift Yeah, but in the meeting he said he's think it over.
17:45 jclift The rpm.t failures (5 of 5) look similar to what I was getting locally when I had bad permissions to the git repo, so mock couldn't execute anything.
17:48 ndevos jclift: if the first test in rpm.t fails (./autogen.sh or ./configure?), all other test will fail too :-/
17:50 ndevos jclift, kkeithley1: I'm working this weekend, and if things stay quiet, I might find some time to look into these build errors
17:50 jclift Doh, of course. :)
17:51 * jclift was looking at it last night
17:51 jclift But the rpm.t stuff itself doesn't really make sense to me
17:52 ndevos jclift: its run from the directory that contains a git/clone and git/cherry-pick of patches,, instead of using that dir, it creates a copy somewhere else
17:52 jclift When stepping through the rpm.t code, I can see it's trying to build stuff, but there's extra code in there using "git describe" in a way that makes no sense to me
17:53 jclift eg: COMMIT=`git describe`
17:53 jclift then
17:54 ndevos jclift: that 'git describe' is done in the current working directory (i.e. original checkout made by Jenkins), after that, a new git repository is created, and the commitid from git-describe is used to get to the same state
17:54 jclift git checkout -b rpm-test $COMMIT
17:54 jclift ndevos: According to the man page, that won't work.  And doesn't seem to in practise.
17:54 ndevos or, maybe its not changing directories, merely creating a new branch?
17:54 jclift "git describe" doesn't generate a commit that can be used in this way
17:55 jclift Yeah, it's definitely creating a new branch.  It _seems_ to always be using git head due to the nonsensical commit that git describe feeds it
17:55 jclift I kind of wonder if the "always using head" is an unintended side effect, but one which is doing what we want in reality :)
17:56 jclift eg always getting git head
17:56 ndevos hmm, I also thought there was a fancy 'git log ...' command and not a 'git describe', but its been a while and I dont have the code open atm
17:56 jclift But I'm not sure.
17:56 jclift Yeah.  I was stepping through it last night manually (eg cut-n-pasting into an open terminal), but then got tired and hit the sack
17:56 jclift Turned out my main prob was dir permissions anyway. ;)
17:57 jclift I think I'll just ignore the regression testing probs for not, and finish getting Glupy working for 3.4.
17:57 ndevos jclift: you can check if ~ndevos/bin/autobuild*.sh on build.gluster.org does it different, I think it just checks out HEAD of the master branch
17:57 jclift s/for not/for now/
17:58 jclift Heh.  I'd forgotten I've got ssh to build.gluster.org.
17:58 ndevos jclift: well, if you dont, I have an internal repo with the scripts too
17:59 jclift Nah, I just ssh'd there, so I think we're ok. :)
18:25 purpleidea joined #gluster-dev
19:19 ndk joined #gluster-dev
21:23 jobewan joined #gluster-dev
21:50 badone joined #gluster-dev
22:07 orion7644 joined #gluster-dev
22:32 xavih_ joined #gluster-dev
22:34 xavih_ joined #gluster-dev
22:59 orion7644 left #gluster-dev
23:32 jobewan joined #gluster-dev

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