Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2015-06-08

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

All times shown according to UTC.

Time Nick Message
00:53 dlambrig left #gluster-dev
01:14 pranithk joined #gluster-dev
03:17 soumya joined #gluster-dev
03:24 shubhendu__ joined #gluster-dev
03:27 johnmark joined #gluster-dev
03:27 nixpanic joined #gluster-dev
03:27 csim joined #gluster-dev
03:28 nixpanic joined #gluster-dev
03:39 nbalacha joined #gluster-dev
03:53 atinmu joined #gluster-dev
03:53 nbalacha joined #gluster-dev
04:07 itisravi joined #gluster-dev
04:12 vimal joined #gluster-dev
04:22 ppai joined #gluster-dev
04:26 kdhananjay joined #gluster-dev
04:28 nkhare joined #gluster-dev
04:30 spandit joined #gluster-dev
04:36 vimal joined #gluster-dev
04:45 kanagaraj joined #gluster-dev
04:52 gem joined #gluster-dev
04:57 jiffin joined #gluster-dev
05:01 sakshi joined #gluster-dev
05:04 hgowtham joined #gluster-dev
05:11 rafi joined #gluster-dev
05:16 poornimag joined #gluster-dev
05:24 Manikandan joined #gluster-dev
05:28 kshlm joined #gluster-dev
05:30 anekkunt joined #gluster-dev
05:34 hagarth joined #gluster-dev
05:35 ashiq joined #gluster-dev
05:42 pppp joined #gluster-dev
05:43 deepakcs joined #gluster-dev
05:53 raghu` joined #gluster-dev
05:55 hchiramm joined #gluster-dev
05:58 schandra joined #gluster-dev
06:00 Gaurav__ joined #gluster-dev
06:03 rjoseph joined #gluster-dev
06:03 atalur joined #gluster-dev
06:05 overclk joined #gluster-dev
06:09 anekkunt kshlm,  Hi .. have a look at this patch http://review.gluster.org/#/c/10197/
06:11 spalai joined #gluster-dev
06:14 Yuan_ joined #gluster-dev
06:24 pranithk joined #gluster-dev
06:26 anekkunt hagarth, Take a look at this patch http://review.gluster.org/#/c/10894/ , it has +1 from Atin and Raghavendra   and dependency patch (http://review.gluster.org/#/c/10197/) also merged
06:34 hagarth anekkunt: will do
06:34 anekkunt hagarth, Thanks
06:36 schandra_ joined #gluster-dev
06:53 anrao joined #gluster-dev
06:54 atinmu kshlm, mind to merge http://review.gluster.org/#/c/11042/ ?
06:54 atinmu kshlm, this is a backport for 3.7
06:54 kshlm What is it?
06:55 kshlm Oh errnos. I'll take a look.
06:56 kshlm atinmu, I doesn't have a NetBSD +1. I'll wait for that.
06:59 atinmu kshlm, I've manually triggered it
06:59 anrao joined #gluster-dev
07:02 hchiramm hagarth,  can u please review/merge this http://review.gluster.org/#/c/10670/  ?
07:03 hchiramm its in release 3.7 branch and backport patch
07:03 hchiramm which is needed for logging patches.
07:06 hagarth hchiramm: will get there after my not-so-short review queue is taken care of :)
07:06 atinmu http://review.gluster.org/#/c/11054/ & 11055 needs to go in mainline, both of them solve spurious failures and few recent regressions have failed because of it
07:06 hchiramm hagarth, ok :)
07:06 atinmu request folks to review and take them in
07:06 hagarth hchiramm: can you also try to get more reviews going before I get to it?
07:06 hchiramm sure..
07:07 krishnan_p joined #gluster-dev
07:08 hchiramm anrao, http://review.gluster.org/#/c/10021/ still merge conflict
07:08 hchiramm http://review.gluster.org/#/c/10532/ same for this
07:09 hchiramm anrao, can u resolve those quickly ?
07:09 anrao hchiramm: I have resolved the merge conflict, but the rfc is throwing internal server error
07:10 hchiramm anrao, ok
07:10 hchiramm anrao++
07:10 glusterbot hchiramm: anrao's karma is now 4
07:13 hchiramm nbalacha++
07:13 glusterbot hchiramm: nbalacha's karma is now 2
07:16 ndevos csim: will we get hyperkitty/mailmain3 on gluster.org?
07:18 csim ndevos: we could, but that's not my priority right now
07:18 csim for now, i wait on package for el7, who need python 3
07:31 ndevos csim: thanks, just asking :)
07:31 hchiramm Manikandan, ping
07:31 glusterbot hchiramm: Please don't naked ping. http://blogs.gnome.org/markmc/2014/02/20/naked-pings/
07:32 Manikandan Humble, pong
07:32 Manikandan Humble, http://etherpad.corp.redhat.com/logging-patches
07:33 soumya joined #gluster-dev
07:33 hchiramm Manikandan++ thanks
07:33 glusterbot hchiramm: Manikandan's karma is now 5
07:33 Manikandan hchiramm, welcome:)
07:33 hchiramm spalai++
07:33 glusterbot hchiramm: spalai's karma is now 2
07:37 tigert hello
07:37 tigert does someone have text content for the 3.7.1 release announcement so I could update the frontpage?
07:56 ndevos hmm, we dont seem to have release-notes for 3.7.1 :-/
07:57 ndevos krishnan_p: did you put release notes for 3.7.1 together? they have not been merged under https://github.com/gluster/glusterfs/tree/release-3.7/doc/release-notes
07:59 ndevos tigert: "Stable releases continue, GlusterFS 3.5.4 is now available" is on the planet, you can promote that ;-)
08:01 ndevos tigert: how about removing "Xen Virtualization on Linux and Solaris" from the planet? I do not think its much gluster related
08:02 tigert ok
08:03 tigert ndevos: you can make pull requests to the planet feeds.yml btw
08:03 * tigert removes that
08:04 tigert removed
08:04 tigert it is going away in a hour
08:04 ndevos tigert++ thanks!
08:04 glusterbot ndevos: tigert's karma is now 11
08:07 tigert ndevos: I need to teach you to make official posts to /news/
08:07 tigert it is easy :)
08:08 csim ndevos: in fact, we could deploy mailman3 on a fedora server , I might explore that
08:09 ndevos csim: sure!
08:12 tigert ndevos: is the 3.5 called "maintenance release" or what?
08:12 tigert pondering a short headline
08:21 owlbot joined #gluster-dev
08:24 ndevos tigert: yeah, I guess "maintenance release" sounds good
08:25 tigert ndevos: site will update soon
08:25 tigert we should figure out some way to pull news to the frontpage via tags or such
08:25 ndevos tigert: awesome!
08:25 tigert maybe just show the latest official blog in /news/
08:27 ndevos not sure how to automate that... I guess additional tag(s) to the blog post would work
08:28 tigert or just show the latest official blog
08:28 ndevos but what makes it "official?"
08:36 tigert it is on gluster.org?
08:37 tigert http://www.gluster.org/news/ < I am talking about this
08:37 tigert this is a "blog"
08:37 tigert we just use it for announcements
08:37 tigert or could
08:37 tigert and we could automatically show the latest post linked from  the frontpage
08:41 hagarth xavih: ping, can you please review http://review.gluster.org/11111 when you get a chance?
08:44 spalai left #gluster-dev
08:46 krishnan_p ndevos, no. I didn't know that release notes were to be merged into our repo.
08:46 krishnan_p ndevos, I sent a summary note on the nature of changes that went in. Is there a format that you would like me to adhere to?
08:48 ndevos tigert: I do not think I have seen /news before, being able to post in my own blog with some additional tags would help me...
08:48 ndevos krishnan_p: I use something like https://github.com/gluster/glusterfs/blob/release-3.5/doc/release-notes/3.5.4.md for the format
08:48 ndevos krishnan_p: but, we dont have rules or guidelines for it, although a common format would be nice :)
08:58 krishnan_p ndevos, let me try to stick to the format.
09:01 krishnan_p ndevos, do you submit a PR for release notes directly to our github repo or do we go via gerrit?
09:01 ndevos krishnan_p: via Gerrit, github just shows the renedered markdown a little nicer
09:01 krishnan_p ndevos, do you have a BZ query that returns the bugs fixed in a given release?
09:02 ndevos krishnan_p: also, I found out that I can push to the branch without going through Gerrit... not sure is that is intentional
09:02 krishnan_p ndevos, hmm, your patch for 3.5.4 release notes didn't have a review.gluster.org url in it's commit log
09:02 krishnan_p ndevos, ah. OK. That explains your commit log :)
09:02 ndevos krishnan_p: something like "git format-patch --format=email --stdout v3.7.0...v3.7.1
09:03 krishnan_p ndevos, ah. nice. For the known issues you must have gone to BZ for the answers. No?
09:03 ndevos ... and then " | grep -e '^BUG' | cut -d: -f 2 | xargs -n1 bugzilla query -b' or something like that
09:04 ndevos krishnan_p: yes, the known issues came from bugzilla and the mailinglists
09:05 krishnan_p ndevos, hmm. I would appreciate if you had the bugzilla command line options handy, at least in your shell's history
09:06 krishnan_p ndevos, wow, that worked. Looks like you type this out every single time :)
09:06 ndevos krishnan_p: I used this to create the list of bugs to close for 3.7.1: it log --format=email v3.7.0...v3.7.1 | grep -i -e '^Bug:' | cut -d ':' -f 2 | sort -u > /tmp/glusterfs-3.7.1-bugs.txt
09:06 * krishnan_p is surpised that bugzilla cli is working out of box for him!
09:07 krishnan_p ndevos++
09:07 glusterbot krishnan_p: ndevos's karma is now 146
09:08 krishnan_p ndevos, Could you review http://review.gluster.com/11095? This is the call_stack to use struct list_head for holding its frames (call_frame_ts).
09:08 ndevos krishnan_p: you can even use the --outputformat option for the bugzilla command, but I normally dont bother with that
09:08 krishnan_p ndevos, this fixes the spurious sparse-self-heal.t failures.
09:08 ndevos krishnan_p: yeah, I was still wrapping my head around that
09:09 krishnan_p ndevos, Hmm. My pitch for this patch was that it made reading STACK_WIND etc easier apart from fixing the spurious failure :)
09:09 krishnan_p ndevos, thanks for looking into the patch.
09:09 krishnan_p ndevos++
09:10 glusterbot krishnan_p: ndevos's karma is now 147
09:11 ndevos krishnan_p++ thanks for doing the release notes :)
09:11 glusterbot ndevos: krishnan_p's karma is now 6
09:18 schandra-remote joined #gluster-dev
09:20 nbalacha joined #gluster-dev
09:20 nbalacha joined #gluster-dev
09:24 tigert ndevos: frontpage now has the 3.5.4 news
09:24 tigert ndevos: I need to check whether the planet supports tags too
09:25 ndevos tigert: nice!
09:25 ndevos tigert: does that come from the planet, or how is it fed into the page?
09:25 hgowtham joined #gluster-dev
09:29 tigert this is a blog engine in the site
09:29 tigert I figured some things are good as news articles on the site itself
09:29 tigert because the planet flows so fast
09:29 tigert stuff gets lost in the stream quickly
09:30 tigert (and for a planet to show something, it has to come from somewhere)
09:30 tigert ndevos: do you have ruby?
09:31 ndevos tigert: ew, no!
09:32 tigert hehe
09:33 tigert ndevos: in a vm! :-)
09:33 tigert its safe that way! :)
09:38 kshlm joined #gluster-dev
09:40 krishnan_p ndevos, how do I identify bugs that were seen before 3.7.1 and didn't make it to 3.7.1?
09:40 krishnan_p ndevos, the first predicate is the one I really need ideas. The second one can be filtered using Fixed in Version, I think.
09:48 kbyrne joined #gluster-dev
09:59 Joe_f joined #gluster-dev
10:06 suliba joined #gluster-dev
10:15 ndevos krishnan_p: bugs filed against version 3.7.* that are not closed?
10:18 ndevos krishnan_p: this should give you a start: https://bugzilla.redhat.com/buglist.cgi?f1=bug_status&amp;f2=version&amp;o1=notequals&amp;o2=regexp&amp;product=GlusterFS&amp;query_format=advanced&amp;v1=CLOSED&amp;v2=^3.7
10:19 ndevos krishnan_p: probably only list the urgent ones, or take the contents from the 3.7.0 release notes
10:20 krishnan_p ndevos, thanks. I will resubmit the release-notes with these included
10:21 pranithk joined #gluster-dev
10:25 krishnan_p ndevos, do we necessarily carry over known issues or could we just point them to look at 3.7.0.md?
10:26 ndevos krishnan_p: I mostly carry them over, they might have been fixed in the update
10:28 krishnan_p ndevos, OK. Let me do that
10:32 kp joined #gluster-dev
10:32 itisravi joined #gluster-dev
10:33 kp ndevos, sorry to bother you again. There are quite a few bugs that didn't make it to 3.7.1, so would it help if I picked Top 10 of urgent and high as part of Known Issues?
10:36 pranithk xavih: Did you get a chance to look at http://review.gluster.org/11111
10:37 xavih pranithk: yes, I posted a couple of comments
10:38 pranithk xavih: Oh cool, let me check :-)
10:39 pranithk xavih: Agree about the 2nd comment. Didn't realize it was there :-)
10:40 pranithk xavih: 1st comment, needs some discussion..
10:41 xavih pranithk: what's the problem ?
10:41 ndevos krishnan_p: yes, something like that, the urgent ones, and/or the ones that popped up on gluster-users recently
10:41 krishnan_p ndevos, OK.
10:41 pranithk xavih: If we take readlink for example.. the softlink could be created while the brick was down. Now when readlink fop is issued it could go to one of the bricks that doesn't have the softlink yet...
10:42 pranithk xavih: I have seen access fop failing with ESTALE sometimes because on that brick the file is not present... when going through some logs a while back...
10:44 xavih pranithk: this means that we probably need to use quorum for these fops also (with a lock and getting size and version before executing them)
10:45 xavih pranithk: there's no other way to be sure of data being really consistent
10:46 pranithk xavih: How can readlink give different results, if the file exists?
10:46 xavih pranithk: doing retries while some error is found will cause unnecessary delays when the actual result must be an error
10:46 pranithk xavih: Just wondering :-)
10:46 atinmu joined #gluster-dev
10:47 pranithk xavih: But it will give very fast reply in case the bricks are all fine (Most probable case)
10:47 xavih pranithk: it could give different answers if bricks were unavailable when changes have been done
10:47 xavih pranithk: how do you know you have asked the good brick ?
10:48 xavih pranithk: what if the a brick was down while the symlink was deleted
10:48 xavih ?
10:48 xavih pranithk: in this case, retrying until one bricks returns success will return the bad answer
10:48 pranithk xavih: readlink comes on gfid. So it either exists or it doesn't. If it does exists then it must have same contents?
10:48 hchiramm ndevos, ping, pm ?
10:49 pranithk xavih: The only way to change contents of softlink is by deleting and re-creating which will lead to different gfid on the softlink...
10:50 pranithk xavih: That is why I thought your implementation was right, except the errnos it wasn't handling..
10:50 kdhananjay joined #gluster-dev
10:50 pranithk xavih: I am not sure if thats the case with access and readdir. Will think about them too
10:50 xavih pranithk: readlink will always receive a gfid ? never a path ?
10:51 pranithk xavih: yes, gfid. This is as per gfs3_readlink_req in rpc/xdr/src/glusterfs3-xdr.h
10:52 pranithk xavih: I think for access and readdir(p), what you said makes sense... We must know which one is the correct brick before winding the fop...
10:54 xavih pranithk: what happens if an xlator sends a readlink request based on a path ? will it fail ?
10:55 hchiramm hagarth, anrao is ready to look at the issue.
10:55 xavih pranithk: a readdir after an opendir will always be directed to a healthy brick. The check is done in opendir
10:56 hagarth hchiramm: great, thanks!
10:57 pranithk xavih: :-( Not any more thanks to my wonderful patch which changed ec_child_select which made sure appropriate flags like EC_FLAG_UPDATE_LOC_PARENT etc are set to consider mask :-(
10:57 hagarth pranithk, itisravi: http://build.gluster.org/job/rackspace-regression-2GB-triggered/10266/consoleFull .. afr/sparse-file-self-heal.t caused a core
10:57 hagarth pranithk, itisravi: mind taking a look in there?
10:57 krishnan_p ndevos, hagarth take a look at http://review.gluster.org/11121 when you get time. This one is the 3.7.1 release notes.
10:58 hagarth krishnan_p: will do
10:58 pranithk kdhananjay, atalur: ^^
10:58 krishnan_p pranithk, isn't the sparse-file-self-heal.t crash due to statedump + stack_reset race?
10:59 pranithk hagarth: ^^
10:59 pranithk krishnan_p: You are right :-)
11:00 pranithk hagarth: krishnan_p sent a patch which I need to revew...
11:00 pranithk xavih: But if we remove those checks also there is a problem... The problem for which we added these flags... Any ideas?
11:01 hagarth pranithk, krishnan_p: great, thanks!
11:01 pranithk xavih: One more beauty is that nfs doesn't send opendir...
11:02 xavih pranithk: yes, I know :(. I'll need to review what's the current state of these problems and see how to solve them
11:03 xavih pranithk: regarding to the patch, your solution for avoiding double unwind is good. We can change the other things when we have a better solution.
11:05 pranithk xavih: Oh you think it is better to do it in two steps... makes sense. I will just address the readdir not doing more winds when the offset is non-zero and then resend the patch
11:05 xavih pranithk: access and readlink also had the same problem...
11:05 xavih pranithk: keep the changes for them also
11:06 xavih pranithk: the critical problem is the crash and this needs to be solved as soon as possible. The other problem requires more thinking...
11:08 pranithk xavih: Sorry I didn't get that. Are you saying there are more problems that we need to address in access/readlink?
11:08 firemanxbr joined #gluster-dev
11:08 atinmu joined #gluster-dev
11:09 xavih pranithk: no, the same problem. The crash could due to double unwind could also happen in access and readlink, isn't it ?
11:09 pranithk xavih: Yeah, the patch has the fixes. Ah! I think you are asking me to keep those changes.
11:10 xavih pranithk: yes :)
11:10 pranithk xavih: will do that :-)
11:12 rafi1 joined #gluster-dev
11:17 hchiramm anrao++ ndevos++
11:17 glusterbot hchiramm: anrao's karma is now 5
11:17 glusterbot hchiramm: ndevos's karma is now 148
11:19 keycto joined #gluster-dev
11:22 rafi joined #gluster-dev
11:31 kdhananjay joined #gluster-dev
11:41 anrao ndevos++
11:41 glusterbot anrao: ndevos's karma is now 149
11:44 kkeithley1 joined #gluster-dev
11:54 rgustafs joined #gluster-dev
11:59 kp joined #gluster-dev
12:00 kp ndevos, I have sent the updated 3.7.1 release notes - http://review.gluster.org/11121.
12:02 rafi1 joined #gluster-dev
12:02 pranithk xavih: Re-sent the patch addressing readdirp change. http://review.gluster.org/#/c/11111/2
12:06 ppai joined #gluster-dev
12:08 kkeithley hchiramm: ping. Please add 3.7.1 to Version, 3.7.2 to Target Milestone in bugzilla. Thanks
12:10 hchiramm kkeithley, sure
12:28 bfoster joined #gluster-dev
12:31 pranithk xavih: http://review.gluster.org/#/c/11111 is resent :-)
12:35 rafi joined #gluster-dev
12:41 kkeithley semiosis has been missing in action for some time. I sent email last week (4 June) and he has not responded. There haven't been any builds in Launchpad for 3.6.3 or 3.7.1; nothing since February.
12:42 kkeithley I've tried submitting builds, but either I don't have the privs or I'm doing it wrong.
12:43 kkeithley Anyone want to venture an opinion about what we should do about Ubuntu pkgs?
12:43 kanagaraj joined #gluster-dev
12:45 kkeithley IOW shall I build them and put them on d.g.o? Or should we wait some more and see semiosis returns?
12:46 ndevos kkeithley: get yourself added to the launchpad team and build the packages there?
12:46 kkeithley I already am added
12:47 kkeithley but either I don't have the privs or I'm doing it wrong.
12:47 kkeithley I've dput a couple source.changes files, with no result.
12:49 ndevos hmm, yes, you're just as "approved" as the "gluster builder" account for all I can see, no idea how launchpad works
12:50 ndevos and gluster-builder did build packages...
12:53 kkeithley I dput the source, launchpad took them. Attempts to resend them are rejected, telling me I've already uploaded that before.
12:54 kkeithley there are no failed builds
12:55 kkeithley or anything else, not even builds that "needs building"
13:27 Joe_f joined #gluster-dev
13:36 shyam joined #gluster-dev
13:42 kdhananjay joined #gluster-dev
13:59 kkeithley oh, well, I was doing it wrong. I seem to have a launchpad build going for 3.6.3
14:02 kkeithley finally
14:13 wushudoin joined #gluster-dev
14:19 lpabon joined #gluster-dev
14:25 ira joined #gluster-dev
14:56 pppp joined #gluster-dev
15:03 pousley joined #gluster-dev
15:13 hagarth joined #gluster-dev
15:34 anrao joined #gluster-dev
15:42 xavih a question about nfs. Some days ago someone said that nfs is killed with signal 9 when glusterd needs to restart it. Is it true ?
15:48 hchiramm kkeithley, the versions are created in bugzilla.
15:48 hchiramm JFY
15:49 hchiramm ah/ JFY/FYI
15:49 hchiramm tdasilva++ thanks for merging that patch
15:49 glusterbot hchiramm: tdasilva's karma is now 2
15:54 nbalacha joined #gluster-dev
15:56 pppp joined #gluster-dev
16:08 jamesp joined #gluster-dev
16:17 ashiq joined #gluster-dev
16:17 Manikandan joined #gluster-dev
16:17 dlambrig1 joined #gluster-dev
16:18 rafi joined #gluster-dev
16:36 hgowtham joined #gluster-dev
16:38 ashiq- joined #gluster-dev
16:59 soumya joined #gluster-dev
16:59 jiffin joined #gluster-dev
17:04 ashiq__ joined #gluster-dev
17:05 rveroy joined #gluster-dev
17:30 Gaurav__ joined #gluster-dev
17:31 jobewan joined #gluster-dev
17:36 ashiq__ joined #gluster-dev
18:45 hchiramm joined #gluster-dev
18:56 pousley joined #gluster-dev
19:35 ndevos oh, Mac OS X again... :-/
21:05 wushudoin| joined #gluster-dev
21:07 badone_ joined #gluster-dev
21:12 wushudoin| joined #gluster-dev
21:13 jbautista- joined #gluster-dev
21:27 jbautista- joined #gluster-dev
21:58 wushudoin| joined #gluster-dev
22:03 wushudoin| joined #gluster-dev

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