Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2015-11-13

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

All times shown according to UTC.

Time Nick Message
00:09 wushudoin joined #gluster-dev
00:18 anrao joined #gluster-dev
01:15 zhangjn joined #gluster-dev
01:19 zhangjn_ joined #gluster-dev
01:52 shyam joined #gluster-dev
01:57 mohan_ joined #gluster-dev
02:07 jvava joined #gluster-dev
02:36 kbyrne joined #gluster-dev
02:45 Chr1st1an joined #gluster-dev
02:45 Chr1st1an joined #gluster-dev
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/
02:48 jvava is it better or worse that using a raid as storage for gluster fs?
02:49 jvava i know hdfs admit not using any raid
02:49 jvava how about gfs
03:06 sakshi joined #gluster-dev
03:30 mohan_ joined #gluster-dev
03:47 overclk joined #gluster-dev
03:56 cholcombe joined #gluster-dev
04:00 bkunal joined #gluster-dev
04:07 bkunal joined #gluster-dev
04:33 cholcombe joined #gluster-dev
04:47 vimal joined #gluster-dev
05:02 kotreshhr joined #gluster-dev
05:04 cholcombe joined #gluster-dev
05:13 mohan_ joined #gluster-dev
05:28 kotreshhr left #gluster-dev
05:38 jvava the official docker file is maintained still now? which seems apparently does NOT work.
05:48 zhangjn joined #gluster-dev
05:49 zhangjn joined #gluster-dev
05:53 overclk joined #gluster-dev
05:59 nishanth joined #gluster-dev
06:01 zhangjn joined #gluster-dev
06:16 zhangjn joined #gluster-dev
06:22 zhangjn joined #gluster-dev
06:41 mjrosenb jvava: gfs meaining glusterfs?
06:48 zhangjn joined #gluster-dev
06:57 zhangjn joined #gluster-dev
07:01 jiffin joined #gluster-dev
07:18 josferna joined #gluster-dev
07:24 sakshi joined #gluster-dev
07:31 sakshi joined #gluster-dev
07:34 zhangjn joined #gluster-dev
08:27 overclk joined #gluster-dev
08:53 mohan joined #gluster-dev
09:05 jvava yes
09:24 mohan joined #gluster-dev
09:28 mjrosenb ok, because I normally assume that gfs is the global file systesm.
09:28 mjrosenb so, where exactly are linktos followed? I assumed dht_lookup?
09:33 zhangjn joined #gluster-dev
09:39 overclk mjrosenb: asking me? if yes, please address the nick, it's easier to follow...
09:44 mjrosenb I'm asking anyone who is around and can answer.
09:48 mjrosenb overclk: ^ Sorry, I'm normally better about irc ettiquite.
09:57 overclk mjrosenb: ok, np. yes, if a lookup() finds a linkto file in the hashed location, the link-to xattr points to the correct location of the file.
09:58 mjrosenb overclk: and that handles both the case of attempting to directly access a file, and getting the contents of the directory that it lives in?
09:59 jvava overclk,  raid will help gluster fs, if not, i will tell my partener
09:59 jvava because i know hdfs does not like it
10:00 jvava raid do help gluster fs?
10:01 overclk mjrosenb: I didn't get your question.
10:01 mjrosenb so, lookup is obviously called when the client attempts to open a file, or chmod a file
10:02 mjrosenb but when I run ls, is lookup also called on every file that is returned?
10:02 overclk mjrosenb: by default, dht send readdir+ down to posix
10:03 overclk mjrosenb: that normally fetches attribute information for all entries that are read
10:03 overclk mjrosenb: and for each entry, the inode is linked in the inode table.
10:04 overclk mjrosenb: but still, the next operation on any entry, needs an explicit lookup() an translators in the stack depend on lookup() to function correctly.
10:08 zhangjn joined #gluster-dev
10:32 jiffin joined #gluster-dev
10:52 mjrosenb uh-oh; /usr/lib/gcc/x86_64-unknown-linux-g​nu/5.2.0/../../../../lib/libfl.so: undefined reference to `yylex'
11:29 lpabon joined #gluster-dev
12:37 ira joined #gluster-dev
12:45 amye joined #gluster-dev
12:46 overclk joined #gluster-dev
14:34 shyam joined #gluster-dev
14:38 mohan joined #gluster-dev
15:12 wushudoin joined #gluster-dev
15:23 jobewan joined #gluster-dev
15:31 amye joined #gluster-dev
16:03 ndevos shyam: are you reviewing and merging http://review.gluster.org/12563 ?
16:04 ndevos thats the libgfapi/unix-domain-socket one
16:24 firemanxbr joined #gluster-dev
16:30 shyam ndevos: heal deamon os started with -s localhost, by glusterd if I am not mistaken, why then can it not start with -s <the IP instead> is what I am wondering
16:30 shyam ndevos: was in a call, saw the above only now...
16:30 ndevos shyam: np :)
16:31 ndevos shyam: glfs-heal is the new libgfapi based heal daemon in 3.7, it has "tcp" and "localhost" hardcoded
16:31 shyam ndevos: Hmmm... :) need to read more
16:32 ndevos shyam: there are other processes started by the gluster-cli too (aux-mount) that have "localhost" hardcoded, a patch has been sent to use the unix-domain-socket there too
16:33 shyam ndevos: I am only seeing if we can avoid an option, so only speaking from that perspective... Is this transparent to the admin/user?
16:33 ndevos shyam: most patches come from ashiq: http://review.gluster.org/#/q/owner:%22​Mohamed+Ashiq+Liyazudeen%22+status:open
16:33 shyam I mean does the admin have to specify this option, or is it automatically done this way? (checking the code)
16:34 shyam ndevos: ok hang on, let me check this, got to take care of another *possibly* urgent task atm (I will merge it)
16:34 ndevos shyam: well, for glfs-heal and other processes/daemons that Gluster automatically starts, it should be transparent to the user - only connecting over the UDS now
16:34 ndevos shyam: sure, take your time
16:47 amye joined #gluster-dev
16:49 overclk joined #gluster-dev
17:26 shaunm joined #gluster-dev
17:32 ira joined #gluster-dev
17:33 ira_ joined #gluster-dev
17:50 cholcombe joined #gluster-dev
18:03 kkeithley er, what's this "unescaped left brace in regex is deprecated, passed through in regex;..."  noise now when I run rfc.sh???
18:15 ndevos kkeithley: I think that comes because of a new perl version, checkpatch.pl
18:15 ndevos kkeithley: I havent looked into it yet, and wont do so this week anymore. Go Perl hacker!
18:16 kkeithley yeah, I thought it might have something to do with moving to f23.
18:16 * ndevos expects so to
18:16 ndevos o
18:17 kkeithley also
18:17 kkeithley backwards compatibility ftw
18:40 csim perl is usually quite good at backward compat
19:30 kkeithley perl regex
20:25 zhangjn joined #gluster-dev
20:50 amye joined #gluster-dev
21:02 cholcombe joined #gluster-dev
22:12 amye joined #gluster-dev
23:24 RedW joined #gluster-dev

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