Camelia, the Perl 6 bug

IRC log for #gluster-dev, 2012-12-19

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

All times shown according to UTC.

Time Nick Message
01:58 maxiz joined #gluster-dev
02:00 bfoster_ joined #gluster-dev
02:01 kkeithley1 joined #gluster-dev
02:01 jdarcy_ joined #gluster-dev
02:08 kshlm|AFK joined #gluster-dev
02:08 kshlm|AFK joined #gluster-dev
02:24 bfoster joined #gluster-dev
02:25 kkeithley joined #gluster-dev
02:28 jdarcy joined #gluster-dev
03:01 sunus joined #gluster-dev
03:31 bulde__ joined #gluster-dev
04:02 bharata joined #gluster-dev
04:43 vpshastry joined #gluster-dev
05:06 hagarth joined #gluster-dev
05:16 mohankumar joined #gluster-dev
05:23 bulde joined #gluster-dev
05:24 sgowda joined #gluster-dev
05:32 mohankumar joined #gluster-dev
05:41 mdarade joined #gluster-dev
05:44 kp_ joined #gluster-dev
06:10 raghu joined #gluster-dev
06:23 mdarade left #gluster-dev
07:04 sunus joined #gluster-dev
07:38 vpshastry1 joined #gluster-dev
07:40 mdarade1 joined #gluster-dev
07:40 mdarade1 left #gluster-dev
08:04 shireesh joined #gluster-dev
08:12 gbrand_ joined #gluster-dev
08:36 mdarade1 joined #gluster-dev
09:17 mohankumar joined #gluster-dev
09:17 vpshastry joined #gluster-dev
09:26 quillo joined #gluster-dev
09:58 puebele1 joined #gluster-dev
10:14 kp_ joined #gluster-dev
10:39 bulde joined #gluster-dev
11:31 blues-man joined #gluster-dev
11:31 blues-man hi guys
11:38 blues-man for an academic work I would better understand which algorithm and theory is behind the geo-replication, I found useful theoretical informations on http://hekafs.org but since I should consider very large amount of files for a geo-replication with Glusterfs, I would being "trusted" by a model
11:39 blues-man is NP? :)
12:16 blues-man joined #gluster-dev
12:18 hagarth joined #gluster-dev
12:43 edward1 joined #gluster-dev
13:40 wushudoin joined #gluster-dev
15:54 sunus joined #gluster-dev
16:28 blues-man joined #gluster-dev
16:31 blues-man hello
16:31 johnmark blues-man: greetings
16:31 blues-man hi johnmark
16:32 semiosis blues-man: geo-rep is done by glusterfs keeping track of what files have changed on the master, then using librsync to do the transfer to the slave
16:33 johnmark blues-man: if you want to have some fun, take a look at how we do geo-rep
16:33 blues-man eheh
16:33 blues-man thank you for answer, I did this morning (GMT+1) but then connection went down
16:33 blues-man so I was wondering if asking in mailing list would be better anyway
16:34 johnmark and for kicks, you may want to check out xsync: https://github.com/avati/xsync
16:34 johnmark blues-man: that's probably your best bet for this kind of question
16:34 blues-man I would understand what happen if master for some reason is unreachable
16:35 johnmark blues-man: the operative term you're looking for is "marker framework" or "marker API" which is how we detect and track files to be synced in geo-rep
16:35 johnmark ah, ok
16:35 johnmark yeah, that's the kind of question you should send to gluster-devel
16:35 blues-man is there a way to ensure consinstency and intergrity of file at all?
16:36 johnmark blues-man: in async replication, there's no way to be absolutely sure
16:36 blues-man because I should make a thesis on glusterfs with the case study of a digitalization of a big library, is would be about 800TB! And so I would studying any consinstent approach
16:36 johnmark our default replication is absolutely consistent
16:37 johnmark but is too slow for replicating over a WAN connection
16:37 johnmark ...generally speaking
16:39 blues-man johnmark, the speed is not a problem for this case, their main problem, and the thesis focus, is to ensure the integrity and consistency of files. Because it is a case study for a possible use of Glusterfs for the Nation Library of Florence, which is going to be digitalized in large part from Google and so the library would to keep those in a secure way
16:39 johnmark interesting
16:39 johnmark I would suggest two things
16:39 johnmark 1. look at AFR, which is our synchronous replication
16:40 ndevos joined #gluster-dev
16:40 johnmark 2. there is a company in Spain working on a sparse replication translator
16:40 johnmark and they might need your help
16:40 johnmark they plan to release a beta of their translator in January
16:40 blues-man I was trying to understand better the theory behind Gluster, and I was look at 3.4 next feature also, the one about Server Quorum that maybe is the one from the theory model (Duvvuri) that could help in this
16:41 johnmark ok
16:41 johnmark you might want to check out the gluster-devel list for those type sof questions
16:41 johnmark or ping jdarcy
16:41 johnmark who hangs out here often
16:43 blues-man I suppose he is the one writing on hekafs.org, as I write before I read some technical documentation over there that helped me to understand better theory behind gluster
16:44 blues-man but know my point is to research the best approach to prove that consistency is guaranteed always, in any worste case
16:45 blues-man thank you johnmark I will look at AFR, about the Spain company are they working on gluster developing also?
16:46 ndevos joined #gluster-dev
16:46 blues-man I'd like to combine the academic research with the aim of the opensource project, trying to contribute somehow
16:48 blues-man It would be nice to say someday Those Operas has proudly saved by Gluster :)
16:49 blues-man *are
17:07 johnmark blues-man: that would be awesome!
17:07 johnmark blues-man: and yes, the company in Spain, Datalab, are working on a Gluster translator to do sparse replication
17:20 hagarth joined #gluster-dev
17:36 mdarade1 left #gluster-dev
17:47 puebele joined #gluster-dev
17:58 blues-man johnmark, yes! :) I just send an email to the devel ml, so I would try to find some info about that company
17:58 blues-man thank you for the help
19:31 gbrand_ joined #gluster-dev
20:19 sghosh joined #gluster-dev
20:57 gbrand__ joined #gluster-dev
20:59 gbrand_ joined #gluster-dev
22:09 inodb_ joined #gluster-dev
22:30 badone joined #gluster-dev
22:30 jdarcy a2: Any thoughts on Amar's patch which failed in regression but then immediately passed smoke?
23:04 avati joined #gluster-dev
23:05 avati ndevos: ping?
23:27 inodb_ left #gluster-dev
23:33 avati joined #gluster-dev
23:41 blues-man joined #gluster-dev
23:41 blues-man hello

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