Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2014-01-17

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

All times shown according to UTC.

Time Nick Message
00:00 purpleidea the funny thing about all this, is i've never tested 3.5.x with puppet-gluster. no idea if anything even works :P
00:00 purpleidea depends on if 3.5 is compatible with 3.4 or not :P
00:05 johnmark lol
00:05 johnmark that's awesome
00:05 johnmark I guess we're about to find outP
00:05 jobewan joined #gluster-dev
00:53 purpleidea johnmark: cool, it all seemed to work great!
00:53 purpleidea just writing it up and then i'll send it your way for review
00:53 purpleidea (10 min)
00:59 johnmark purpleidea: awesome - thanks :)
01:16 purpleidea johnmark: let me know if you have any changes or comments before i publish: http://ttboj.wordpress.com/?p=71​3&shareadraft=52d8846fef3ec
01:17 purpleidea kkeithley_: code is published too in case you see any issue with: https://github.com/purpleidea/puppet-gluster/c​ommit/30392fd0cb4e2bd0e39faea83915bfe8a6574bbc
01:18 bala joined #gluster-dev
01:19 johnmark purpleidea: ah, very very nice
01:19 johnmark way cool - let it fly
01:24 purpleidea johnmark: i hate to use sentences like this but: "do you have a hash tag for glusterfest" ?
01:33 johnmark lol
01:33 johnmark I used it on twitter
01:33 johnmark purpleidea: ^^^
01:33 johnmark so yes
01:33 johnmark just #glusterfest
01:35 purpleidea johnmark: https://twitter.com/purpleide​a/statuses/423990127610712064
01:36 johnmark woohoo
01:36 johnmark sweetness
01:36 purpleidea all done.
01:36 johnmark is my weakness
01:36 purpleidea did you see the git commit comment ?
01:55 bharata-rao joined #gluster-dev
02:02 johnmark ha no
02:13 purpleidea johnmark: https://github.com/purpleidea/puppet-glu​ster/commit/30392fd0cb4e2bd0e39faea8391
02:14 purpleidea going for pizza bbl if you need anything or if anyone has problems with the puppet/vagrant stuff, just ping me
02:28 johnmark purpleidea: lol... I thought you already had a polo :)
02:47 ilbot3 joined #gluster-dev
02:47 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/
03:14 kanagaraj joined #gluster-dev
03:17 bala joined #gluster-dev
03:32 shubhendu joined #gluster-dev
03:45 itisravi joined #gluster-dev
03:50 ppai joined #gluster-dev
03:54 [o__o]2 joined #gluster-dev
04:11 kanagaraj joined #gluster-dev
04:18 kanagaraj joined #gluster-dev
04:28 purpleidea johnmark: nope. i've got a glusterfs sticker :) it has a picture of you on it apparently
04:28 johnmark haha - ok, well, there's a polo for you. Not sure how you didn't get one :/
04:29 * johnmark adds it to the list of things to send to purpleidea
04:29 purpleidea no worries! size medium would be great!
04:48 bala joined #gluster-dev
04:51 spandit joined #gluster-dev
05:06 kdhananjay joined #gluster-dev
05:07 lalatenduM joined #gluster-dev
05:32 pk1 joined #gluster-dev
05:39 ira joined #gluster-dev
05:46 shyam joined #gluster-dev
05:50 mohankumar joined #gluster-dev
05:51 ndarshan joined #gluster-dev
05:55 raghu joined #gluster-dev
06:24 bala joined #gluster-dev
06:41 bulde joined #gluster-dev
06:45 hagarth joined #gluster-dev
06:51 bulde joined #gluster-dev
07:17 aravindavk joined #gluster-dev
07:22 bala joined #gluster-dev
07:22 kanagaraj_ joined #gluster-dev
07:26 kanagaraj joined #gluster-dev
07:34 kanagaraj joined #gluster-dev
08:09 kanagaraj joined #gluster-dev
08:24 ababu joined #gluster-dev
08:40 kanagaraj joined #gluster-dev
08:49 ababu joined #gluster-dev
08:53 aravindavk joined #gluster-dev
08:53 shubhendu joined #gluster-dev
08:54 ndarshan joined #gluster-dev
09:00 shyam joined #gluster-dev
09:10 bala joined #gluster-dev
09:46 mohankumar__ joined #gluster-dev
10:13 kanagaraj joined #gluster-dev
10:19 hagarth bharata-rao: ping
10:26 bharata-rao hagarth, hi
10:30 hagarth joined #gluster-dev
10:31 ndarshan joined #gluster-dev
10:31 shubhendu joined #gluster-dev
10:31 aravindavk joined #gluster-dev
10:31 hagarth bharata-rao: regarding the sector size issue with xfs .. IIRC, we need the sector size value to be small enough so that writes from qemu go through fine right?
10:32 hagarth when the caching mode is  none?
10:32 bala joined #gluster-dev
10:38 bharata-rao hagarth, generally yes, but there are a few points
10:39 bharata-rao hagarth, qemu block layer and seabios don't know about 4kb sectors
10:39 bharata-rao hagarth, so you can't use a disk with 4kb sector size as rootdisk
10:39 hagarth bharata-rao: right
10:40 bharata-rao hagarth, but you can use it as a data disk by explicitly specifying the sector size on qemu cmd line
10:40 hagarth bharata-rao: ok
10:41 bharata-rao hagarth, also iirc you can workaround this for rootdisk by explicitly specifying the format of the VM image on qemu cmdline since that results in qemu not "open"ing that image to check the format
10:41 bharata-rao hagarth, ^ need to verify that once though
10:42 bharata-rao hagarth, and all this is in the context of cache=none which results in O_DIRECT
10:43 hagarth bharata-rao: right.. other modes should be fine
10:46 bharata-rao hagarth, yes I think so
10:47 hagarth bharata-rao: cool, thanks!
10:47 bharata-rao hagarth, hey wc
10:50 ppai joined #gluster-dev
10:54 purpleidea hagarth: hey btw, i did that repo.pp action item thing of yours from 2 weeks ago... https://github.com/purpleidea/puppet-gluster/c​ommit/30392fd0cb4e2bd0e39faea83915bfe8a6574bbc i then tested the puppet-gluster+vagrant stuff with 3.5.0beta and it seems to work great! good job on 3.5 and thank you for not breaking the interfaces i use :)
10:54 purpleidea hagarth: article about the details here: https://ttboj.wordpress.com/2014/01/16​/testing-glusterfs-during-glusterfest/
10:56 hagarth purpleidea: cool, glad that it's working for you :)
10:57 purpleidea yep! :)
11:35 kanagaraj joined #gluster-dev
11:47 ira joined #gluster-dev
12:04 itisravi joined #gluster-dev
12:06 edward1 joined #gluster-dev
13:23 yinyin joined #gluster-dev
13:43 bala joined #gluster-dev
13:52 hagarth joined #gluster-dev
14:24 lalatenduM hagarth, ping
14:40 jobewan joined #gluster-dev
15:42 lalatenduM joined #gluster-dev
15:49 yinyin joined #gluster-dev
16:23 lpabon joined #gluster-dev
16:59 hagarth lalatenduM: pong
17:01 lalatenduM hagarth, was looking for tips to debug the rebalance bug I raised
17:31 hagarth lalatenduM: sorry was afk, does the problem happen even with a static data set?
17:31 hagarth lalatenduM: i.e. without any activity from clients
17:39 lalatenduM hagarth, yes I think
17:39 lalatenduM hagarth, when I tried the force option , there was no I/O but it still gave errors
17:39 lalatenduM i mean I saw the issue
17:42 hagarth I see .. a simple rebalance seems to be working for me
17:43 lalatenduM hagarth, may be it is because of the directory depth i.e. multi level directory
17:44 lalatenduM hagarth, you can check the script I used to run the IO, I have mentioned it in the bug
17:44 hagarth lalatenduM: possible.. checking out the logs in the bug now
17:50 hagarth lalatenduM: 2f218e1335d5fdab0b41716cc5c8976b20c367f6 is the patch which has altered the behavior of rebalance
17:50 * lalatenduM looking
17:51 hagarth http://review.gluster.org/6341 is the patch URL
17:53 lalatenduM hagarth, awesome , you got it within minutes
17:55 lalatenduM hagarth, yup the error is introduced in that patch
17:55 hagarth lalatenduM: the check for incrementing failure count should not be done for directories
17:55 hagarth it should be done only for files
17:55 hagarth all error messages in the logs seem to be for directories
17:59 lalatenduM hagarth, I think I understand what u r sayin, ENODATA is ok for directories (for dht) (as I remember from sishir's talk) , it should not be counted as failure
17:59 lalatenduM as looks can also fix the layout or somethin
18:00 lalatenduM s/looks/lookups/
18:00 hagarth right .. I will ask pk to send a patch for this
18:01 lalatenduM hagarth, awesome :),
18:07 hagarth lalatenduM: hmm, I seem to have a different theory now .. it is not due to directories
18:08 hagarth but a manifestation of how errno and syncop work together
18:08 lalatenduM hagarth, :)
18:08 lalatenduM hagarth, sounds like a bigger problem
18:09 hagarth lalatenduM: http://lists.nongnu.org/archive/html​/gluster-devel/2013-12/msg00036.html
18:10 * lalatenduM looking
18:10 lalatenduM hagarth, I had filed that bug :) for which pk had sent the patch
18:11 hagarth lalatenduM: yeah, you seem to be a specialist in this area now :)
18:11 lalatenduM hagarth, LOL
18:14 hagarth lalatenduM: it is interesting to note that pranith's interim fix was to turn off optimization for the same function
18:15 hagarth for some reason it doesn't seem to be helping
18:15 lalatenduM hagarth, this change went to master, did it also went to 3.5 branch
18:15 lalatenduM hagarth, I think the interim patch was working isn't it
18:15 hagarth lalatenduM: good point, it is not present in release-3.5
18:16 hagarth so we just have to get a backport of the interim patch
18:16 lalatenduM hagarth, do you have the commit id of the interim patch handy
18:18 hagarth lalatenduM: http://review.gluster.org/6481
18:19 lalatenduM hagarth, awesome :)
18:21 hagarth lalatenduM: shall be off now, glad that we figured out the answer to your bug pretty soon :)
18:22 lalatenduM hagarth, yup, thanks to you :) gud night
18:22 hagarth lalatenduM: thanks to you too! good night
18:35 Technicool joined #gluster-dev
19:25 jobewan joined #gluster-dev
19:25 jobewan joined #gluster-dev
23:31 jobewan joined #gluster-dev

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