Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2014-07-07

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

All times shown according to UTC.

Time Nick Message
01:33 bala joined #gluster-dev
03:05 hagarth joined #gluster-dev
03:58 spandit joined #gluster-dev
03:59 kanagaraj joined #gluster-dev
03:59 itisravi joined #gluster-dev
04:07 kdhananjay joined #gluster-dev
04:08 ppai joined #gluster-dev
04:31 ndarshan joined #gluster-dev
04:32 vipulnayyar joined #gluster-dev
04:42 shubhendu joined #gluster-dev
04:49 nishanth joined #gluster-dev
04:55 shubhendu joined #gluster-dev
05:26 bala1 joined #gluster-dev
05:26 jv115 joined #gluster-dev
05:27 krishnan_p joined #gluster-dev
05:28 vimal joined #gluster-dev
05:32 hagarth joined #gluster-dev
05:35 lalatenduM joined #gluster-dev
05:47 skoduri joined #gluster-dev
05:57 vpshastry joined #gluster-dev
06:05 kshlm joined #gluster-dev
06:12 skoduri joined #gluster-dev
06:13 rgustafs joined #gluster-dev
06:31 raghu joined #gluster-dev
06:50 hagarth joined #gluster-dev
07:03 atinmu joined #gluster-dev
07:35 pranithk joined #gluster-dev
07:35 pranithk ndevos: Vijay wanted geo-rep patches to be merged asap. So
07:36 ndevos pranithk: sure, if there is a reason, push me :D
07:37 pranithk ndevos: yep!
07:37 pranithk JoeJulian: ping
07:37 pranithk semiosis: ping
07:37 pranithk Who is online when either Joe or semiosis are online?
07:37 ndevos pranithk: it helps if others do the review +2, I only check the very basics then
07:38 ndevos pranithk: I mostly am, or kkeithley, or some others guys in the US
07:39 hagarth pranithk: JoeJulian is on the west coast and semiosis is on the east coast - if that helps to find an overlapping TZ
07:39 pranithk ndevos: Basically debian packaging is missing /usr/bin/glfsheal binary. I need to get them to look at it
07:39 pranithk ndevos: Sure I'll get the +2 don't worry :-)
07:41 ndevos pranithk|lunch: did you file a bug in launchpad(?) agains the .deb packaging?
08:15 pranithk ndevos: no. Peter3 in #gluster reported it
08:15 pranithk ndevos: I found the root cause to be that the binary is missing in .deb package
08:16 ndevos pranithk: okay, so semiosis and/or JoeJulian should be aware of the issue already? Got a link to the bug?
08:16 pranithk ndevos: gimme a sec
08:19 pranithk ndevos: bz1113778
08:20 pranithk ndevos: Damn! it doesn't give bot reply :-(. Let me try https://bugzilla.redhat.co​m/show_bug.cgi?id=1113778
08:20 glusterbot Bug 1113778: medium, unspecified, ---, pkarampu, ASSIGNED , gluster volume heal info keep reports "Volume heal failed"
08:20 ndevos pranithk: different bots everywhere, bug 1113778
08:20 glusterbot Bug https://bugzilla.redhat.com:​443/show_bug.cgi?id=1113778 medium, unspecified, ---, pkarampu, ASSIGNED , gluster volume heal info keep reports "Volume heal failed"
08:20 pranithk ndevos: grrrrr :-)
08:23 ndevos pranithk: dont you want to assign that bug to semiosis?
08:23 pranithk ndevos: I don't know his id. Do you?
08:24 pranithk ndevos: I tried sending him email on the mail-id I know of. But he didn't respond
08:24 ndevos pranithk: re-assigned to him now, but please add a clear note on what he should be looking at :)
08:24 ndevos ah, no, it failed!?
08:26 pranithk ndevos: Sure I will add the note.
08:28 ndevos pranithk: strange, none of the email addresses I have (or are in .mailcap in the glusterfs sources) seem to work :-/
08:29 pranithk ndevos: what to do :-(
08:30 itisravi ndevos: hi Neils, http://review.gluster.org/#/c/8229/ still fails the bug-822830.t regression inspite of pranith's patch (http://review.gluster.org/#/c/8235/) that you merged.
08:30 glusterbot Title: Gerrit Code Review (at review.gluster.org)
08:31 ndevos pranithk: say 'semiosis' 3x and he might appear?
08:32 pranithk ndevos: :-).
08:33 itisravi ndevos: the test passes on my local machine, wondering if the build system isn't making the build off the current HEAD
08:33 ndevos pranithk: the email you have from him on CC seems to be the one he used most recently for other bugs too, so that should be good...
08:34 pranithk ndevos: cool
08:36 ndevos itisravi: to check change http://review.gluster.org/#/c/8229/, I go to the webui and select 'download', 'checkout', 'ssh' and copy the git command
08:36 glusterbot Title: Gerrit Code Review (at review.gluster.org)
08:37 ndevos itisravi: I do not think http://review.gluster.org/#/c/8235 has been included in release-3.5 though, checking what that change actually is
08:37 glusterbot Title: Gerrit Code Review (at review.gluster.org)
08:38 ndevos itisravi: ah, right, http://review.gluster.org/8235 is included, but indeed not in the branch that seems to get tested...
08:38 glusterbot Title: Gerrit Code Review (at review.gluster.org)
08:38 itisravi ndevos: yup
08:38 ndevos or, at least not in the branch that gerrit suggests me to check-out
08:39 itisravi ndevos: should I trigger another buid and see?
08:39 ndevos itisravi: not yet, lets see the console log and check what git-commits are used there
08:40 itisravi ndevos: okay
08:43 itisravi ndevos: it says GIT_COMMIT=0b6afe45a355350​4f51ca3ed6cc3b89ee3db5720 but I can't find any commit on master or release-3.5 with that git commit id
08:44 ndevos itisravi: that is the commit of your patch :)
08:44 ndevos itisravi: and indeed, I come to the same conclusion as you, the regression test did not pick the current state of release-3.5, but the one you based the patch on
08:44 itisravi ndevos: oh shucks you're right
08:45 ndevos itisravi: maybe the git cherry-pick script for rackspace-regression-2GB-triggered is different than the normal regression test
08:46 itisravi ndevos: looks like that seems to be the case
08:46 ndevos itisravi: can you start a normal regression test for your change, and see if that works out well
08:46 ndevos itisravi: for changes (and issues) to the rackspace-regression-2GB-triggered job, you should get in touch with JustinClift
08:47 itisravi ndevos: will do. also, would you know why we are triggering netbsd-smoke tests on 3.5 branch?
08:47 itisravi ndevos: ok, will check with him
08:48 ndevos itisravi: I think we'd like to see NetBSD improve, I'm happy to accept (simple) patches in 3.5 for that
08:50 itisravi ndevos: mhm..those tests are failing as well, but due to compile errors. Maybe I should ask Emmanuel Dreyfus for that.
08:51 itisravi ndevos: one more thing, I am not able to add you as a reviewer on gerrit using your redhat ID :)
08:52 ndevos itisravi: try typing my full name, start with: Niels
08:52 ndevos itisravi: it shows my private email address, but gerrit sends it to my @redhat.com anyway...
08:53 * ndevos thinks his gerrit accounts are a little messed up
08:53 itisravi ndevos: i see the fedoraproject account and nixpanic.net.
08:53 ndevos itisravi: nixpanic.net is the one
08:54 itisravi ndevos: okay, it said "don't-use@nixpanic" so I didn't :)
08:54 ndevos itisravi: but, for release-3.5 you do not really need that, I've setup a filter/preference and receive all release-3.5 emails already
08:54 itisravi ndevos: ah okay
08:54 ndevos itisravi: ah, strange, that is indeed the wrong one then :-/
08:54 ndevos well, and the fedoraproject.org one isnt correct either
08:56 itisravi ndevos: oh. anyway since you've setup a filter I think it is okay.
08:56 ndevos hchiramm__: help? could you delete an account in Gerrit for me? the account with devos@fedoraproject.org is broken, account-id 1000051 is the correct+working one
09:00 JustinClift Ahhh, that's an interesting distinction between the triggered job and the normal rackspace regression
09:00 JustinClift itisravi ndevos: The triggered one lets Gerrit decide what to checkout and test
09:01 JustinClift So yeah, it's completely possible it's using the git branch that was submitted, rather than merging the specific CR onto HEAD of the approach branch
09:02 JustinClift itisravi ndevos: With the "normal" rackspace regression one (which copied its approach from the old non-rackspace job), the script itself does the checkout and merge
09:02 itisravi JustinClift: so for now maybe I should update my local branch before hitting rfc.sh
09:02 JustinClift itisravi: Always a good idea :)
09:02 itisravi JustinClift: got it :)
09:03 ndarshan joined #gluster-dev
09:03 aravindavk joined #gluster-dev
09:05 ndevos itisravi: rebasing the change before ./rfc.sh makes it more time consuming to review updated patches, I tend to compare differences between patch versions and a rebase makes that very messy
09:05 ndevos JustinClift: could you update the triggered regression test to use the same checkout/cherry-pick as the manual regression one?
09:10 JustinClift ndevos: With a bunch of effort, probably
09:11 JustinClift ndevos: Probably need to blow away the git repo, and re-check it out manually using the same code as the other ones
09:11 JustinClift I guess this is why the triggered version never throws up merge conflicts ;)
09:12 JustinClift ndevos: Ahhh, I'd not seen comparing differences between patch versions actually work before
09:12 JustinClift ndevos: I guess the ones I'd looked at with that had all rebased between the submission.  But yeah, being able to compare between patch revisions would make things a bunch easier.
09:14 ndevos JustinClift: yeah, checking differences between patch versions is really nice
09:15 ndevos JustinClift: I'm not sure how much the regression test scripts in Jenkins change, but having the (almost) same script would surely be good :)
09:17 JustinClift The triggered one was a lot simpler, as I was able to cut out 60% of the scripting.
09:17 JustinClift (which was to do with the manual git checkout)
09:17 JustinClift Guess I'd better start mucking around with a v2 someone soon
09:18 * JustinClift was so hoping to move off of the whole regression testing stuff by now ;)
09:21 pranithk JustinClift: I don't think you should do the re-submissions of the regression jobs.
09:22 pranithk JustinClift: Also let me know if I can help you out in anyway.
09:29 JustinClift pranithk: This isn't clear to me: "I don't think you should do the re-submissions of the regression jobs."
09:29 JustinClift eg wut? :)
09:30 JustinClift On a side not, having the env variables displayed alphabetically in those console logs does look a lot nicer
09:30 JustinClift s/side not/side note/
09:30 JustinClift :)
09:32 pranithk JustinClift: Don't do manual resubmission by yourself. It is tedious. Let people take responsibility for their own patch.
09:32 pranithk JustinClift: Most of the folks in Redhat know Amar's username/password to re-trigger failed jobs
09:33 JustinClift Ahhh
09:33 JustinClift k, np :)
09:36 JustinClift ndevos pranithk: Btw, I've left that hanging regression test alone on purpose, so Manu can log in and investigate it while it's still hung.
09:36 JustinClift It does block one slave, but I don't think the queue will get too out of shape due to it
09:38 pranithk JustinClift: agree!
09:40 ndevos pranithk: ah, and I was wondering why amar started all those regression tests :D
09:41 pranithk ndevos: Damn! I think I outed so many people :-D
09:44 JustinClift Heh
09:44 JustinClift ndevos: Yeah, I've emailed Amar a few times asking "wtf did you run that again for?" and never got a response
09:44 JustinClift Now I know what ... ;)
09:44 ndevos JustinClift: lol!
09:44 JustinClift s/what/why/
09:44 JustinClift But yeah, not important.  It's obviously working for ppl
09:46 JustinClift ndevos pranithk: What do you guys think of having a gluster CLI option that outputs the various run time dir locations for things?
09:46 JustinClift eg for things like /var/run/glusterfs
09:46 JustinClift eg config file location, log file location, that kind of thing
09:46 JustinClift Sort of similar to what pkg-config foes
09:47 JustinClift s/foes/does/
09:47 JustinClift There are a couple of new CR's in gerrit that really want to compile .c code as part of their tests
09:48 JustinClift And atm there's no "easy to use" way of finding out where the installed libraries on on a system, in a cross platform way
09:48 JustinClift We could muck about with .in files, but that seems non-optimal
09:48 pranithk JustinClift: I think there are already some commands liek that
09:48 ndevos JustinClift: like "httpd -V"?
09:48 JustinClift pranithk: Oh that would be great
09:49 JustinClift ndevos: I don't know what httpd -V goes ;)
09:49 * JustinClift goes and finds a httpd box
09:49 JustinClift ndevos: Ahh yeah, something like that
09:49 ndevos pranithk: if its there, cool, but why does 'glusterfs --help' not show it?
09:50 pranithk ndevos: they are hidden commands
09:50 pranithk ndevos: check cli/src/cli-cmd-system.c
09:50 * JustinClift rolls eyes
09:50 ndevos pranithk: hidden is not very helpful :D
09:50 JustinClift tx
09:50 JustinClift Lets unhide them ;)
09:50 pranithk ndevos: Not all of them.
09:51 pranithk ndevos: I donno. We can bring it up in the next meeting?
09:51 JustinClift Sure
09:51 JustinClift ... or even discuss on gluster-devel ?
09:52 ndevos pranithk: cli/src/cli-cmd-system.c are for the gluster-cli, thats not really what I'd use for it, and those command are hidden away just fine there :)
09:52 ndevos JustinClift: for a "httpd -V" equivalent, you can just file a bug
09:54 JustinClift ndevos: Gotcha
09:54 JustinClift Will do
09:55 ndevos JustinClift: btw, what would you want to do with it?
09:55 ndevos JustinClift: oh, and when you file that bug, set the EasyFix keyword too ;)
10:00 JustinClift ndevos: Good thinging
10:00 JustinClift ndevos: For this stuff, it'll make it easier to integrate with stuff
10:01 JustinClift ndevos: Also some of the proposed Gerrit changes are failing atm, because people want to do things like compile bits of .c to test specific functionality, and they're having trouble finding the right way to do it
10:01 ndevos JustinClift: EasyFix bugs seem to get some attraction, at least I've added a topic of "send your first patch" to the tasks of newly hired colleagues ;)
10:02 JustinClift :)
10:02 JustinClift ndevos: A thing like this that can show where the base dir is, will help people figure out where their dependant library is installed
10:03 JustinClift So that can use -L/path/ or whatever
10:08 bala1 joined #gluster-dev
10:20 aravindavk joined #gluster-dev
10:23 hagarth joined #gluster-dev
10:43 vpshastry joined #gluster-dev
10:50 jcastillo the EasyFix flag is an amazing way to encourage learning
10:51 pranithk jcastillo: hey! I guess you are new to gluster? saw some patches and your email id suggest you are from redhat
10:51 pranithk s/suggest/suggests
10:52 edward1 joined #gluster-dev
11:08 aravindavk joined #gluster-dev
11:13 ppai joined #gluster-dev
11:20 hagarth joined #gluster-dev
11:20 shyam joined #gluster-dev
11:20 skoduri joined #gluster-dev
11:23 vpshastry joined #gluster-dev
11:27 lkoranda joined #gluster-dev
11:32 kkeithley joined #gluster-dev
12:11 aravindavk joined #gluster-dev
12:25 itisravi_ joined #gluster-dev
12:28 systemonkey2 joined #gluster-dev
12:28 glusterbot` joined #gluster-dev
12:31 glusterbot joined #gluster-dev
12:37 suliba_ joined #gluster-dev
12:57 lalatenduM ndevos, I want to work on another easy fix bug too, but I forgot the bugzilla query for it
12:58 ndevos lalatenduM: bugzilla query --product=GlusterFS --keywords=EasyFix --bug_status=NEW
12:58 lalatenduM ndevos, thanks
12:58 ndevos lalatenduM: there must be a link to the query in some emails too...
12:59 lalatenduM ndevos, yeah I was looking for that
12:59 lalatenduM never mind I will search it
12:59 ndevos lalatenduM: oh, and maybe in the wiki, not sure if we started an EasyFix page already
13:00 lalatenduM ndevos, nope we dont have, I think we should start a page
13:00 lalatenduM will do it
13:00 skoduri joined #gluster-dev
13:02 pranithk joined #gluster-dev
13:02 pranithk JoeJulian: ping
13:02 pranithk semiosis: ping
13:02 pranithk ndevos: ping
13:03 ndevos pranithk: again?
13:03 pranithk ndevos: again what?
13:03 ndevos pranithk: you?
13:03 pranithk ndevos: I am not understanding your question :-(
13:04 pranithk ndevos: I pinged you to ask for http://review.gluster.org/8251 http://review.gluster.org/8250 to be merged...
13:04 glusterbot Title: Gerrit Code Review (at review.gluster.org)
13:04 pranithk kkeithley: ping
13:04 ndevos pranithk: okay, will try to review them later today
13:05 pranithk ndevos: cool sir!
13:05 pranithk jcastillo: ping
13:05 kkeithley pranithk: pong
13:05 pranithk kkeithley: Corvidtech found one more issue, I found it to be a leak of client_t. Would love your review for http://review.gluster.org/#/c/8247/
13:05 glusterbot Title: Gerrit Code Review (at review.gluster.org)
13:06 ndevos pranithk: care to review http://review.gluster.org/8193 too then?
13:06 glusterbot Title: Gerrit Code Review (at review.gluster.org)
13:06 kkeithley pranithk: I'll take a look
13:06 pranithk kkeithley: thanks :-)
13:08 jcastillo pranithk: pong, hello!
13:08 pranithk ndevos: gave +2 merge it merge it meger it ;-)
13:09 pranithk jcastillo: hey!, I saw your patches and from your email id found that you are also from redhat. Curious to know which team you work for.
13:09 ndevos pranithk: hehe, done!
13:10 jcastillo pranithk: gss :)
13:10 pranithk jcastillo: oh! nice :-)
13:10 pranithk jcastillo: which country?
13:10 jcastillo pranithk: based in UK, Farnborough
13:10 ndevos lalatenduM: btw, https://bugzilla.redhat.com/buglist.cgi?bug_statu​s=NEW&keywords=EasyFix&product=GlusterFS would do it too
13:10 glusterbot Title: Bug List (at bugzilla.redhat.com)
13:11 pranithk jcastillo: Cool.
13:11 lalatenduM ndevos, thanks, yeah got it and assigned one bug to myself
13:11 ndevos pranithk: jcastillo is in the same team as I am :D
13:11 pranithk ndevos: ah! got it.
13:12 jcastillo yeah, I'm basically ndevos's minion :D
13:12 pranithk ndevos: hehe :-)
13:12 hagarth joined #gluster-dev
13:13 ndevos haha, but only what's Gluster or RHS related :)
13:13 lalatenduM ndevos, pranithk creating a wiki page for easyfix bugs and will link it from Developers page
13:13 jcastillo lalatenduM: I have a patch almost ready for 1092158, how far away did you get?
13:13 pranithk lalatenduM: Would love that. Thanks for doing it. I was supposed to do it...
13:13 ndevos lalatenduM: cool, thanks!
13:14 lalatenduM jcastillo, just picked it , how come we always pick the same bug to fix :)
13:14 lalatenduM jcastillo, assign it to ur name and make a comment that u r working on it
13:15 jcastillo lalatenduM: I guess because the list of EasyFix is too small for now :D
13:15 jcastillo lalatenduM: yeah doing that now, sorry man
13:15 lalatenduM jcastillo, no prob :)
13:15 pranithk jcastillo: Good to see that the EasyFix idea is getting some activity :-)
13:16 lalatenduM pranithk, ndevos , should I mention that if anybody wantto work on a bug , he/she should assign it to his/her name first?
13:16 ndevos jcastillo, lalatenduM: when you pick an EasyFix you want to work on, assign it to yourself and move it to ASSIGNED, that should prevent confusion :)
13:16 pranithk lalatenduM: Yes. To avoid collisions
13:16 lalatenduM ndevos, :)
13:16 ndevos pranithk: we need more EasyFix bugs
13:16 pranithk ndevos: There is a triage happening on wednesday right?
13:17 ndevos pranithk: wasnt it friday?
13:17 ndevos pranithk: well, teh meeting I mean?
13:17 pranithk ndevos: probably. I forgot :-(
13:17 vpshastry joined #gluster-dev
13:17 pranithk ndevos: We can take care of these things there.
13:17 ndevos pranithk: yes, we should :)
13:18 pranithk ndevos: cool
13:21 pranithk JustinClift: ping
13:23 kkeithley pranithk: are you going to backport http://review.gluster.org/8247 to 3.5?
13:23 glusterbot Title: Gerrit Code Review (at review.gluster.org)
13:23 pranithk kkeithley: As soon as I learn that it is merged :-)
13:24 pranithk kkeithley: Thanks a lot for the review :-)
13:27 kkeithley yw
13:42 pranithk ndevos: ping
13:42 ndevos pranithk: pong!
13:44 ndevos pranithk: "We would love a bug." ? You can't get enough of those, right?
13:44 kanagaraj joined #gluster-dev
13:44 pranithk ndevos: :-D
13:44 pranithk ndevos: After installing I did not install wireshark. Gurdeep on gluster-users sent wireshark capture file. Do you have time to analyse that one and check if it is coming from self-heal-daemon.
13:44 pranithk ndevos: After installing F20!
13:44 lalatenduM pranithk, ndevos please review http://www.gluster.org/community/do​cumentation/index.php/EasyFix_Bugs :)
13:44 glusterbot Title: EasyFix Bugs - GlusterDocumentation (at www.gluster.org)
13:45 pranithk lalatenduM: On it
13:45 lalatenduM ndevos, I copied few text from the mail (which you have written)
13:45 ndevos pranithk: well, uhm, not sure of we can see that from the capture file alone...
13:45 ndevos lalatenduM: okay, thanks for getting that started :)
13:45 pranithk ndevos: check if frame->root->pid is LARGE number.
13:46 pranithk lalatenduM: Good job :-)
13:46 lalatenduM pranithk, thanks
13:46 pranithk ndevos: self-heal fops come with pid -1, so it is easy to figure out.
13:47 ndevos pranithk: right, but was a capture attached to his email?
13:48 ndevos pranithk: ah, yet an other email
13:48 tdasilva joined #gluster-dev
13:50 ndevos pranithk: PID=14927, UID=48
13:50 pranithk ndevos: :-(
13:51 pranithk ndevos: Thanks for the help!
13:52 ndevos pranithk: http://paste.fedoraproject.org/115993/40474114
13:52 glusterbot Title: #115993 Fedora Project Pastebin (at paste.fedoraproject.org)
13:53 ndevos pranithk: around line 93 is the PID/UID/GID
13:54 pranithk ndevos: Found it :-)
13:54 hchiramm__ ndevos :)
14:01 lalatenduM ndevos, regarding https://bugzilla.redhat.co​m/show_bug.cgi?id=1073111, you have sent one patch and then JustinClift has another one. is it ok if I take it
14:01 glusterbot Bug 1073111: low, unspecified, ---, jclift, NEW , %post install warning for glusterfs-server that it can't find /etc/init.d/glusterfsd (on EL6)
14:04 vpshastry joined #gluster-dev
14:05 hagarth pranithk: did you check what the pid mentioned by ndevos corresponds to?
14:06 lpabon joined #gluster-dev
14:08 hagarth pranithk: that will give you an indication of the application that can be causing the lookups
14:08 pranithk hagarth: isn't it pid of the applicaiton process?
14:08 hagarth pranithk: right, that is what we need to find out
14:08 ndevos lalatenduM: sure, go for it
14:09 hagarth the application logic might be triggering lookups
14:09 pranithk hagarth: He was saying that even though no operations were going on the network is being used.
14:09 hagarth pranithk: those might be operations which he does not expect from the application
14:09 hagarth pranithk: strace -f of the application can help determine if the application is actually doing something on the mount point.
14:10 hagarth I hope he has updatedb.conf updated to prune fuse.glusterfs .. else a crawl might trigger several lookups
14:10 pranithk hagarth: Yes. Got it
14:12 pranithk just responded
14:13 ndevos pranithk, hagarth: the strange things is that there are only LOOKUPs in the attached tcpdump, no READ, READDIR or something that suggests crawling
14:14 pranithk ndevos: interesting. Lets wait for his response?
14:14 hagarth ndevos: client side caching can prevent fops like stat, read etc. from appearing over the wire
14:15 pranithk ndevos: Looks lot like stats. md-cache can prevent stats going out
14:16 ndevos hagarth, pranithk: ah, cache might be something, the lookups are mainly done on directories or image files
14:16 hagarth ndevos: my hunch is that his application is updatedb (trying to re-validate its db) ;)
14:18 ndevos hagarth: thats well possible, from the capture, all .jpg files have exactly 4 lookups, thats an amazing coincidence
14:19 hagarth I am awaiting his response before going out for a run ;)
14:20 pranithk hagarth: Since you are waiting anyway, why not update -1, -2 messages on gerrit? ;-)
14:21 hagarth pranithk: after taking a backup of r.g.o ;)
14:21 kkeithley joined #gluster-dev
14:22 hagarth looks like he has gone slow, bbl
14:22 ndevos hagarth: have a good run :)
14:22 pranithk what is r.g.o?
14:23 * ndevos guesses: review.gluster.org
14:23 pranithk ndevos: got it
14:23 pranithk ndevos: he responded
14:25 ndevos hmm, bug 1099460 seems to get more interest...
14:25 glusterbot Bug https://bugzilla.redhat.com:​443/show_bug.cgi?id=1099460 high, high, ---, ndevos, ASSIGNED , file locks are not released within an acceptable time when a fuse-client uncleanly disconnects
14:25 pranithk ndevos: Yeah!
14:27 ndevos pranithk: what did he respond?
14:28 pranithk ndevos: he gave new capture files in his latest response
14:31 pranithk ndevos: ok wireshark installed on my machine as well. I am also taking a look
14:31 pranithk ndevos: big file :-)
14:33 ndevos pranithk: you call 1.4 MB 'big'?
14:33 * ndevos is used to get GBs of captures...
14:33 pranithk ndevos: :-)
14:34 pranithk ndevos: Some problem with my installation of wireshark :-(... debugging that
14:34 ndevos pranithk: what do you need from the capture?
14:36 pranithk ndevos: Same info. What is the process which is doing the fops
14:37 ndevos pranithk: PIDs: http://paste.fedoraproject.org/116010/74385214
14:37 glusterbot Title: #116010 Fedora Project Pastebin (at paste.fedoraproject.org)
14:38 ndevos pranithk: thats the output of this: tshark -r glus.pcap -T fields -e rpc.auth.pid "rpc.msgtyp == 0" | sort | uniq -c
14:39 deepakcs joined #gluster-dev
14:39 pranithk ndevos: looking
14:40 pranithk ndevos: I don't see any self-heal daemon again
14:40 ndevos pranithk: fwiw, it's all UID=48 again, it may be interesting to know what user that would be
14:41 pranithk ndevos: I really feel you should respond. You are doing the debugging after all.
14:41 pranithk ndevos: Just give him the link to find what are those pids
14:42 ndevos pranithk: I think you're a great proxy :D
14:42 pranithk ndevos: you bad boy! I'll do that then
14:45 pranithk ndevos: done!
14:46 ndevos pranithk: https://access.redhat.com/documentation/en-U​S/Red_Hat_Enterprise_Linux/5/html/Deployment​_Guide/s1-users-groups-standard-users.html suggests that UID=48 is apache :)
14:46 glusterbot Title: 37.3. Standard Users (at access.redhat.com)
14:49 pranithk ndevos: Respond you bad boy
14:49 ndevos pranithk: if you explained how to figure out the user, he'll find it, we dont know what distro he's running, I'm only guessing :)
14:50 pranithk ndevos: Ah got it. I just asked him to figure out who is UID 48 as well
14:51 ndevos pranithk: cool
14:53 pranithk ndevos: alright dude. Cya. Logging off.
14:53 ndevos pranithk: argh, http://build.gluster.org/job/rack​space-regression-2GB/543/console failed again due to bug-1112559.t ... and I can't merge it because of the Verified -1 :-/
14:53 glusterbot Title: rackspace-regression-2GB #543 Console [Jenkins] (at build.gluster.org)
14:53 ndevos pranithk: sure, have a good evening!
14:54 pranithk ndevos: I already sent a mail to the dev about this issue. Lets see.
14:54 ndevos pranithk: yeah, I've seen it, thanks for bugging them ;)
14:54 pranithk ndevos: Whats the use. He didn't respond at all the previous time :-(
14:55 ndevos pranithk: I'm glad you're trying!
14:56 JoeJulian pranithk: pong
14:56 pranithk ndevos: but it is discouraging :-(.
14:56 pranithk JoeJulian: finally!
14:56 pranithk JoeJulian: I almost left a minute back
14:57 pranithk JoeJulian: 3.5 ubuntu debs don't have /usr/bin/glfsheal binary
14:57 ndevos pranithk: yes, I know the feeling, but don't give up, I very much appreciate it :)
14:57 pranithk JoeJulian: 'heal info' command uses this binary to fetch the required info.
14:58 pranithk JoeJulian: We were desparately trying to contact semiosis. But he is not responding either :-(
14:58 ndevos JoeJulian: that would be bug 1113778
14:58 glusterbot Bug https://bugzilla.redhat.com:​443/show_bug.cgi?id=1113778 medium, unspecified, ---, pkarampu, ASSIGNED , gluster volume heal info keep reports "Volume heal failed"
14:58 pranithk ndevos: Impeccable timing sir
14:58 JoeJulian I'll send him a text.
14:59 ndevos pranithk: I can be your proxy too :)
14:59 pranithk ndevos: :-). It is nice if people respond. That's all I am saying.
15:00 pranithk JoeJulian: But I wanted to know if it is something I could have fixed at the time of sending the patch. I changed glusterfs.spec.in to include the binary for packaging. Is there something more we need to do?
15:00 pranithk JoeJulian: For debs I mean
15:00 shyam joined #gluster-dev
15:00 JoeJulian pranithk: Since this is always up on my screen, proxied, logged, and everything else, feel free to just leave your message. Naked pings are frustrating if you're gone by the time I see it. :P
15:00 ndevos pranithk: oh, definitely, responding to questions/pings/emails is what keeps me most busy all day long
15:00 johnmark pranithk: I recommend public shaming
15:01 ndevos johnmark: so, after emails to the mailinglist dont work, a blog post with a "WANTED" picture?
15:01 johnmark lol possibly?
15:01 JoeJulian I'm not sure why that's not packaged. I'll take a look at how the package building works and let you know.
15:01 johnmark when was hte last time we heard fro mhim?
15:02 JoeJulian semiosis was last seen in #gluster 3 days, 17 hours, 2 minutes, and 34 seconds ago
15:02 pranithk johnmark: Lets not do anything drastic here. I can go and talk to him tomorrow.
15:03 ndevos johnmark: pranithk and I were not looking for semiosis ;)
15:03 pranithk ndevos: +1
15:03 johnmark ndevos: ah, didn't realize that :)
15:03 johnmark ok
15:03 johnmark never mind
15:03 ndevos well, we were too, but not for the shaming
15:03 johnmark heh heh ok
15:04 JoeJulian Awe... :( It's always fun to shame semiosis
15:04 pranithk JoeJulian: you guys! :-)
15:09 pranithk JoeJulian: you are running mysql on gluster?
15:10 JoeJulian Yes. I have been for years.
15:12 pranithk JoeJulian: wow! I never knew anyone who did that. purpleidea on glusterusers said you presented about it?
15:13 JoeJulian Yep. I did some preliminary testing and innodb split so that each file is on a different dht subvolume was faster than local on rackspace VMs.
15:14 pranithk JoeJulian: slides?
15:14 pranithk JoeJulian: sounds really cool by the way :-)
15:15 JoeJulian I'll have to see if I still have them. They might have only been on a dead hard drive. I never did finish testing that like I wanted to though.
15:17 pranithk JoeJulian: hmm... cool.
15:20 pranithk ndevos: ping
15:21 ndevos pranithk: oh, still here?
15:21 pranithk ndevos: yes
15:21 * JoeJulian takes away pranithk's ping button.
15:21 pranithk JoeJulian: :-)
15:22 ndevos thank you JoeJulian!
15:22 ndevos wait, didnt glusterbot know about naked pings?
15:22 pranithk ndevos: I was wondering if there is a way to figure out how long are we taking to move a community bug from NEW to ASSIGNED
15:22 pranithk ndevos: Wasn't that in #gluster?
15:24 ndevos pranithk: hmm, you could check the history of a bug, but thats quite a manual way of doing things... not sure if it is possible to create a report for that
15:24 pranithk ndevos: So I was watching http://www.youtube.com/watch?v=SzGi1DfbZMI
15:24 glusterbot Title: 2011 Djangocon US keynote David Eaves - Open Source Community Building - YouTube (at www.youtube.com)
15:24 pranithk ndevos: He suggests creating incentives to change the behavior of the communities.
15:25 pranithk ndevos: This would come in handy for us IMO
15:27 ndevos pranithk: oh, I think stats like that are interesting, bugzilla has https://bugzilla.redhat.com/report.cgi , but I'm not sure yet how to use it
15:27 glusterbot Title: Reporting and Charting Kitchen (at bugzilla.redhat.com)
15:27 JoeJulian mp add " ping$" "echo $nick: Please don't naked ping. http://blogs.gnome.org/mark​mc/2014/02/20/naked-pings/"
15:27 JoeJulian @mp add " ping$" "echo $nick: Please don't naked ping. http://blogs.gnome.org/mark​mc/2014/02/20/naked-pings/"
15:27 glusterbot JoeJulian: Error: You don't have the admin; channel,op capability. If you think that you should have this capability, be sure that you are identified before trying again. The 'whoami' command can tell you if you're identified.
15:27 JoeJulian @mp add " ping$" "echo $nick: Please don't naked ping. http://blogs.gnome.org/mark​mc/2014/02/20/naked-pings/"
15:27 glusterbot JoeJulian: The operation succeeded.
15:28 pranithk JoeJulian: ping
15:28 glusterbot pranithk: Please don't naked ping. http://blogs.gnome.org/mark​mc/2014/02/20/naked-pings/
15:28 pranithk JoeJulian: it works!
15:28 JoeJulian Well, we'll see if it works...
15:28 JoeJulian Whether it works or not is kind-of up to you. :P
15:29 kkeithley too many rules
15:29 kkeithley my religion/faith requires that I only do naked pings
15:29 JoeJulian lol
15:30 purpleidea pranithk, JoeJulian: i too want the slides :)
15:30 JoeJulian It does say "please".
15:30 kkeithley And I'm prepared to take it all the way to SCOTUS
15:30 JoeJulian purpleidea: It's good to want things. It builds character.
15:30 purpleidea hehe
15:31 purpleidea kkeithley: i think there's a $DEITY variable in the standard env you can set for those types of comments
15:31 JoeJulian Just set it in your default hiera data.
15:33 pranithk ndevos: Do you know anyone who knows these kinds of report generations? We *must* give the latency of moving a bug from NEW to ASSIGNED
15:33 nishanth joined #gluster-dev
15:36 pranithk ndevos: we can talk about it tomorrow. Cya for today.
15:36 ndevos pranithk: I assume some managers would know, I have to think about it a little who would be able to help us with that
15:36 pranithk ndevos: cool.
15:36 ndevos pranithk: sure, have a good evening!
15:36 pranithk JoeJulian: please don't forget to ping semiosis
15:37 JoeJulian Will do, goodnight.
16:27 johnmark I know of a few communities that do resolution time stats
16:27 johnmark I think cloudstack is one of them
16:28 * johnmark checks
16:32 _Bryan_ joined #gluster-dev
16:49 kkeithley JustinClift: ping.  For mellanox connectx3 vpi cards, which firmware do I want from  http://www.mellanox.com/page​/firmware_table_ConnectX3IB ?
16:49 JustinClift kkeithley: What's the card model?
16:50 JustinClift kkeithley: If the card is installed in a server, you should be able to get its PSID directly from it
16:50 JustinClift eg there some util to query the thing.  It'll tell you the PSID value.  Look that up in the table on that web page
16:50 kkeithley CONNECTX-3 VPI ADPT CARD-SGL PT QSFP QDR IB
16:51 JustinClift Yeah, that's not helpful. ;)
16:51 kkeithley looking for something in the dmesg
16:51 JustinClift kkeithley: Is it plugged into a box, or do you have the card physically in your hard or ?
16:51 JustinClift Ahhh
16:51 kkeithley it's in the box
16:51 kkeithley boxes
16:51 kkeithley they're in the boxes
16:51 JustinClift k, let me try and remember the command that'll do it
16:51 JustinClift (it's been a long time)
16:53 JustinClift Hmmm, probably the flint command
16:53 JustinClift Which OS on the box atm?
16:54 JustinClift kkeithley: ^^
16:54 kkeithley they're new cards. They shouldn't need firmware actually. I'm looking for Mlx drivers. I was told to use Mlx's drivers
16:54 kkeithley RHEL6
16:54 JustinClift Oh
16:54 JustinClift It's weird to be told to use the Mlx drivers instead of the RHEL provided ones.  Any idea why?
16:54 kkeithley Mlx's OFED
16:55 JustinClift Yeah, I know what you're meaning.  It's just unusual as they're not officially supported by RH (apparently)
16:56 kkeithley IIRC, RHEL drivers aren't maintained any more. That's what the guy who has helped me previously told me. (But he's on PTO atm)
16:57 JustinClift k.  Normally the thing to do is install the "Infiniband Support" group, but yeah, if they're not maintained any more then yeah, go for the Mellanox OFED ones.
16:57 kkeithley Previously he put MLNX_OFED_LINUX-2.1-1.0.0.23610 stack, but then someone reprovisioned the machines out from under me
16:57 JustinClift Ahhh
16:57 JustinClift So you're wanting the download URL for that?
16:57 * JustinClift can find it
16:57 MacWinner joined #gluster-dev
16:57 kkeithley yeah
16:57 kkeithley or newer
16:57 JustinClift np
16:57 kkeithley I saw 2.2 mentioned
16:58 JustinClift Yeah, 2.2 should be ok
16:58 JustinClift You were probably looking on this page:
16:58 JustinClift http://www.mellanox.com/page/products_dyn?p​roduct_family=26&mtag=linux_sw_drivers
16:58 glusterbot Title: Mellanox Products: Mellanox OpenFabrics Enterprise Distribution for Linux (MLNX_OFED) (at www.mellanox.com)
16:58 JustinClift Guessing this: http://www.mellanox.com/page/mlnx_ofed_eula?mta​g=linux_sw_drivers&mrequest=downloads&m​type=ofed&mver=MLNX_OFED-2.2-1.0.1&mnam​e=MLNX_OFED_LINUX-2.2-1.0.1-rhel6.5-x86_64.tgz
16:58 glusterbot Title: Mellanox Technologies: MLNX_OFED EULA (at www.mellanox.com)
16:58 JustinClift (that's specific to rhel 6.5)
16:58 kkeithley didn't see the Linux SW/Drivers link when I was there previously
16:59 JoeJulian @mp add "(\S+): \+\+" "$1++"
16:59 glusterbot JoeJulian: The operation succeeded.
16:59 JoeJulian @mp add "(\S+)\+\+" "$1++"
16:59 glusterbot JoeJulian: The operation succeeded.
17:00 JoeJulian @mp add "(\S+)\-\-" "$1--"
17:00 glusterbot JoeJulian: The operation succeeded.
17:00 JoeJulian @mp add "(\S+): \-\-" "$1--"
17:00 glusterbot JoeJulian: The operation succeeded.
17:00 kkeithley @karma
17:00 JustinClift kkeithley: Near the bottom of the page there's a bit with several tabs.  "Benefits" "MLNX_OFED 2.x" "MLNX_OFED 1.6.3-X" "Archive"
17:00 glusterbot kkeithley: Highest karma: "joejulian" (5), "semiosis" (2), and "hagarth" (2). Lowest karma: "Humble" (1), "kkeithley" (1), and "hchiramm_ " (1). You (kkeithley) are ranked 9 out of 14.
17:00 kkeithley JusttinClift: yup, downloading now
17:00 JustinClift Cool. :)
17:14 jobewan joined #gluster-dev
17:30 _Bryan_ joined #gluster-dev
17:38 vimal joined #gluster-dev
17:44 hchiramm_ joined #gluster-dev
17:46 shyam joined #gluster-dev
18:31 kkeithley JustinClift: ping, still there?
19:01 kkeithley JustinClift: never mind, I'll catch you tomorrow
19:52 JustinClift kkeithley: Oops, wasn't watching IRC
19:52 JustinClift kkeithley: Btw, what's the output from ibv_devinfo look like on the boxes with those IB cards in them?
22:17 systemonkey joined #gluster-dev
22:47 badone joined #gluster-dev

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