Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2014-05-26

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

All times shown according to UTC.

Time Nick Message
00:38 yinyin_ joined #gluster-dev
01:49 awheeler joined #gluster-dev
02:04 bala joined #gluster-dev
03:13 krishnan_p joined #gluster-dev
03:27 kshlm joined #gluster-dev
03:38 shubhendu joined #gluster-dev
03:50 itisravi joined #gluster-dev
04:03 skoduri joined #gluster-dev
04:15 ppai joined #gluster-dev
04:22 skoduri1 joined #gluster-dev
04:24 shubhendu joined #gluster-dev
04:26 skoduri joined #gluster-dev
04:33 kdhananjay joined #gluster-dev
04:37 spandit joined #gluster-dev
04:49 bala joined #gluster-dev
04:56 ndarshan joined #gluster-dev
05:03 bala joined #gluster-dev
05:14 _Bryan_ joined #gluster-dev
05:14 nishanth joined #gluster-dev
05:25 lalatenduM joined #gluster-dev
05:36 hagarth joined #gluster-dev
05:41 itisravi_ joined #gluster-dev
05:42 hagarth1 joined #gluster-dev
05:42 itisravi_ joined #gluster-dev
05:57 itisravi joined #gluster-dev
05:57 hagarth joined #gluster-dev
06:02 hagarth1 joined #gluster-dev
06:09 bala1 joined #gluster-dev
06:18 kanagaraj joined #gluster-dev
06:21 hagarth joined #gluster-dev
06:32 kshlm joined #gluster-dev
06:39 skoduri joined #gluster-dev
06:47 kaushal_ joined #gluster-dev
07:40 edward1 joined #gluster-dev
07:41 ndevos hmm, anyone knows whats up with build.gluster.org?
07:50 hagarth ndevos: had to reboot it again. Humble - would it be possible to check why build.gluster.org is getting into an unresponsive state?
07:50 ndevos hagarth: :-/ thanks
07:52 Humble hagarth, I dont think I have the access ..
07:53 hagarth Humble: please send across your public key to me.
07:53 ndevos hagarth: btw, the v3.5.1beta tag got pushed to github and the forge, did you do that or is that automated?
07:54 hagarth ndevos: there is an automated job which does a git push periodically from gerrit to forge & github. Maybe that pushed the tag. I have not done it manually.
07:56 ndevos hagarth: oh, wait, the forge does not have it, but github does (or did?)
07:56 ndevos yes, github has it: https://github.com/gluster/glusterfs/releases/tag/v3.5.1beta
07:56 glusterbot Title: Release v3.5.1beta · gluster/glusterfs · GitHub (at github.com)
07:56 ndevos hagarth: it seems it still needs pushing to the forge...
07:59 hagarth ndevos: shall check
07:59 ndevos hagarth: it's not urgent :)
08:00 hagarth ndevos: shall check = mental note for doing it later ;)
08:00 ndevos hehe
08:14 rgustafs joined #gluster-dev
08:24 awheeler joined #gluster-dev
08:31 scuttle_ joined #gluster-dev
08:49 WangFeng joined #gluster-dev
08:51 WangFeng left #gluster-dev
08:58 krishnan_p joined #gluster-dev
09:37 pranithk joined #gluster-dev
09:38 pranithk ndevos: xavih: I think I found one of the reasons why softlinks of directories are missing
09:39 xavih pranithk: really ?
09:39 ndevos pranithk: wohoo!
09:39 pranithk ndevos: xavih: wait let me get the lin
09:39 pranithk link
09:40 pranithk ndevos: xavih: https://bugzilla.redhat.com/show_bug.cgi?id=1101143#c1
09:40 glusterbot Bug 1101143: medium, medium, ---, vbellur, ASSIGNED , link file under .glusterfs directory not found for a directory
09:40 pranithk ndevos: When is 3.5 release?
09:40 pranithk ndevos: I want avati to look at it and ok it then pull it down
09:41 ndevos pranithk: I've done a 3.5.1beta this weekend, hopefully some testers confirm that bugs are fixed and a release can be made later this week
09:42 ndevos pranithk: if there are urgent things to fix, make sure to file a bug and add glusterfs-3.5.1 in the 'blocks' field
09:42 pranithk ndevos: xavih: I am working on the test script at the moment. my goal is to post the patch on master today.
09:42 pranithk ndevos: Hmm... Let me see what xavih says
09:42 pranithk ndevos: xavih: brb
09:44 xavih pranithk: I'm looking at the bug. However I think Avati is on holiday
09:47 aravindavk joined #gluster-dev
09:48 xavih pranithk: it makes sense, and it really can be the cause of the problem we detected in one of our customers because they did many directory renames, and some of them could have been done with some brick down...
09:49 xavih pranithk: I'll try to reproduce it
09:49 pranithk xavih: its very easy. I am going to send the patch today. You can check that.
09:50 xavih pranithk: ok
09:50 xavih pranithk: good catch :)
09:50 pranithk xavih: rename directory which already has some files. after healing wait for some 10 minutes so that janitor will do the magic.
09:50 pranithk xavih: Thanks to you! for bringing it to my notice.
09:51 pranithk xavih: I could never believe that such a think happened.
09:55 ndevos pranithk: is there any tool that checks the correctness of a .glusterfs directory?
09:55 ndevos pranithk: like, a gfid should be a symlink to a directory, or the link-count should be 2+
09:56 ndevos and, maybe verify the gfid xattr with the name under .glusterfs/ too?
10:04 pranithk ndevos: xavih: Not that I know of
10:05 pranithk ndevos: I actually found one intersting issue in posix_istat() call
10:05 ndevos pranithk: do you think we should we provide such a fsck.glusterfs-brick script?
10:05 ndevos pranithk: you did?
10:06 pranithk ndevos: It just copies the gfid that is passed to the function instead of doing a getxattr
10:06 ndevos pranithk: so, it assumes the xattr matches the gfid-link?
10:09 pranithk ndevos: it seems to :-(
10:11 ndevos pranithk: well, its faster... and if all goes right, it should not be a problem...
10:11 pranithk ndevos: yep
10:24 awheeler joined #gluster-dev
10:25 hagarth joined #gluster-dev
10:29 xavih pranithk: it happens as you have told :)
10:30 xavih ndevos: I'm not aware of anything like that
10:31 ndevos xavih, pranithk: just sent an email to gluster-devel, if you think something like that might be good, please respond by email :)
10:32 itisravi joined #gluster-dev
10:33 xavih ndevos: I've just seen it
10:58 tdasilva joined #gluster-dev
11:01 skoduri joined #gluster-dev
11:03 edward1 joined #gluster-dev
11:15 bala1 joined #gluster-dev
11:24 ppai joined #gluster-dev
11:37 pranithk ndevos: Sure... I will do that
12:03 pranithk left #gluster-dev
12:35 lalatenduM joined #gluster-dev
13:08 wushudoin joined #gluster-dev
13:09 hagarth joined #gluster-dev
13:16 wushudoin joined #gluster-dev
13:19 rgustafs joined #gluster-dev
13:28 lalatenduM ndevos, ping, http://review.gluster.org/#/c/7362/2
13:28 glusterbot Title: Gerrit Code Review (at review.gluster.org)
13:29 lalatenduM ndevos, need little help to understand the change
13:31 ndevos lalatenduM: whats the question?
13:33 lalatenduM ndevos, lets start with change in the file glusterd-hooks.c, what is it try to do
13:33 lalatenduM ndevos, it return 0 if there are .rpmsave files present right?
13:35 ndevos lalatenduM: it returns 0 if the filename ends with '.rpmsave'
13:36 ndevos lalatenduM: well, in fact, it probably should check for *.rpmnew too :-/
13:38 lalatenduM ndevos, ok, how does it solve the issue if it returns 0 if a file present with ending rpmsave ?
13:39 ndevos lalatenduM: it is not solving the issue, but preventing execution of the hook when a backup file was created (the backup of the hook script would be name $ORIGNAME.rpmsave)
13:41 ndevos lalatenduM: so, in the end, only one hook with the real filename should get executed, scripts named *.rpmsave or *.rpmnew should be skipped
13:42 lalatenduM ndevos, got it. I dont think we create rpmnew files, let me check
13:44 lalatenduM ndevos, I may be wrong, the old changes are there in the spec file any more
13:45 lalatenduM ndevos, that is my next doubt, how does the spec file decides to rename old hook scripts as rpmsave
13:49 ndevos lalatenduM: %config(noreplace) would create .rpmnew files instead of .rpmsave
13:50 ndevos lalatenduM: those hook directories are like a public API, we don't know what users put in there, or how they package them
13:51 lalatenduM ndevos, ok. we are using %config without noreplace , does it mean it will replace the existing hook script and make the old as rpmsave?
13:52 ndevos lalatenduM: yes, exactly that
13:57 lalatenduM ndevos, what happens when we have a old hook rpmsave file and corresponding hook script, after upgrade we will have previously in use hook script as rpmsave and the new one?
13:57 lalatenduM and the old rpmsave will be deleted?
14:04 ndevos lalatenduM: can you rephrase that?
14:05 lalatenduM ndevos, with the change , what will be the behavior if we already have hookscripts ending with rpmsave
14:07 ndevos lalatenduM: we should never execute hook scripts that end in .rpmsave (or .rpmnew), hopefully the scripts contain some logic that they don't change the settings again (or whatever the script would do)
14:09 lalatenduM ndevos, my question is little different. lets say, we have a.sh and a.sh.rpmsave . now after update, we will have a.sh and a.sh.rpmsave right? not a.sh, a.sh.rpmsave and a.sh.rpmsave.rpmsave
14:10 ndevos lalatenduM: ah, right, no, there is no a.sh.rpmsave.rpmsave (thats not a RPM standard)
14:12 lalatenduM ndevos, cool :) thanks for answering all my question :)
14:12 ndevos lalatenduM: np!
14:21 ilbot3 joined #gluster-dev
14:21 Topic for #gluster-dev is now Gluster Development Channel - http://gluster.org | For general chat go to #gluster | Patches - http://review.gluster.org/ | Channel Logs - https://botbot.me/freenode/gluster-dev/ & http://irclog.perlgeek.de/gluster-dev/
14:27 yinyin_ joined #gluster-dev
14:36 awheeler joined #gluster-dev
14:56 yinyin_ joined #gluster-dev
15:35 pranithk joined #gluster-dev
15:40 kanagaraj joined #gluster-dev
15:43 pranithk left #gluster-dev
15:48 awheeler joined #gluster-dev
16:27 wushudoin joined #gluster-dev
16:29 bala joined #gluster-dev
17:18 lalatenduM joined #gluster-dev
17:53 rgustafs joined #gluster-dev
18:15 bala joined #gluster-dev
20:13 edward1 joined #gluster-dev
20:14 edward1 joined #gluster-dev
20:16 wushudoin joined #gluster-dev
21:39 wushudoin joined #gluster-dev
22:10 edward2 joined #gluster-dev

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