Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2015-01-27

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

All times shown according to UTC.

Time Nick Message
02:02 badone joined #gluster-dev
02:37 bala joined #gluster-dev
02:53 pp joined #gluster-dev
02:56 spandit joined #gluster-dev
03:22 bala joined #gluster-dev
03:47 hagarth joined #gluster-dev
04:04 jamesc|2 joined #gluster-dev
04:09 ndarshan joined #gluster-dev
04:15 kanagaraj joined #gluster-dev
04:16 shubhendu joined #gluster-dev
04:23 spandit joined #gluster-dev
04:24 spandit joined #gluster-dev
04:26 atinmu joined #gluster-dev
04:28 kshlm joined #gluster-dev
04:28 kshlm joined #gluster-dev
04:31 lalatenduM joined #gluster-dev
04:34 anoopcs joined #gluster-dev
04:41 jiffin joined #gluster-dev
04:41 gem joined #gluster-dev
04:45 pranithk joined #gluster-dev
04:55 hagarth joined #gluster-dev
04:56 Manikandan joined #gluster-dev
05:03 sakshi joined #gluster-dev
05:09 rafi1 joined #gluster-dev
05:17 ppai joined #gluster-dev
05:20 pp joined #gluster-dev
05:28 kdhananjay joined #gluster-dev
05:30 nkhare joined #gluster-dev
05:39 hagarth joined #gluster-dev
05:48 soumya_ joined #gluster-dev
05:56 itisravi joined #gluster-dev
05:58 atalur joined #gluster-dev
06:07 soumya_ joined #gluster-dev
06:14 nishanth joined #gluster-dev
06:19 hchiramm joined #gluster-dev
06:20 hagarth joined #gluster-dev
06:29 rjoseph|afk joined #gluster-dev
06:37 hchiramm joined #gluster-dev
06:43 pranithk joined #gluster-dev
07:09 hchiramm joined #gluster-dev
07:21 atinmu joined #gluster-dev
07:23 ppai joined #gluster-dev
07:32 bala joined #gluster-dev
08:01 hagarth joined #gluster-dev
08:02 ppai joined #gluster-dev
08:20 atinmu joined #gluster-dev
08:37 kdhananjay joined #gluster-dev
08:49 hchiramm joined #gluster-dev
08:54 atalur joined #gluster-dev
08:58 lalatenduM ndevos, I dont see the python scripts for changing bug status in https://forge.gluster.org/bugzap​pers/nightly-builds/trees/master, look like I am looking at a wrong repo
09:00 ndevos lalatenduM: remind me on what scripts?
09:01 ppai joined #gluster-dev
09:01 overclk joined #gluster-dev
09:03 ndevos lalatenduM: ah, maybe you mean some of the scripts I did not publish (except for pastebin) yet?
09:05 aravindavk joined #gluster-dev
09:08 lalatenduM ndevos, sorry was away
09:09 lalatenduM ndevos, the scripts which you run after a upstream release to change bugs state
09:09 lalatenduM ndevos, I mean bugzilla scripts
09:10 nishanth joined #gluster-dev
09:16 atalur joined #gluster-dev
09:30 anoopcs joined #gluster-dev
09:39 itisravi_ joined #gluster-dev
09:52 shubhendu joined #gluster-dev
09:52 shaunm joined #gluster-dev
09:53 bala joined #gluster-dev
09:54 ndevos is Gerrit misbehaving for others as well?
09:54 ndevos pages dont seem to load for me, or only partially :-/
09:56 pranithk ndevos: worked fine for me just now for http://review.gluster.org/#/c/9482/3/api/src/glfs.c
09:57 ndevos pranithk: hmm, I only get the header/footer and no contents :-/
09:57 * ndevos goes off and installs all fedora updates
09:59 * ndevos biab.
10:01 hchiramm its working for me .
10:04 rjoseph|afk joined #gluster-dev
10:10 ndevos pranithk: oh, lol, you were actually looking at the same change I amt trying to open :)
10:11 pranithk ndevos: I knew you are interested in that, so I opened that change :-D
10:13 ndevos pranithk: :P so, what does the lk_owner structure contain? and do you think it makes sense to pass that to libgfapi-users?
10:14 pranithk ndevos: let me check
10:18 pranithk ndevos: patch looks fine
10:18 ndevos pranithk: okay, but do we want/need that functionality exposed?
10:19 * ndevos was going to check if the now-void-pointer-not-lk_owner-structure-anymore also contains a size in the function calls
10:20 pranithk ndevos: Soumya things we need that. Not sure why. If we think it is needed the patch is definitely doing that
10:20 pranithk ndevos: I am trying to understand why we are believing void* is definitely a gf_lkowner_t
10:20 hagarth joined #gluster-dev
10:21 ndevos pranithk: I understand why soumya_ would like it, it makes implementing lock in ganesha easier
10:21 ndevos pranithk: but, I also think ganesha is responsible for the locking, and it should be transparent to any posix-like filesystem
10:22 * ndevos blames GPFS for some of the workarounds in ganesha
10:23 soumya_ ndevos, no you got it wrong...I am not going enable that 'lock_supported_owner' right away even if this patch goes in :P
10:23 soumya_ ndevos, this is one thing which we recently came across while trying to understand how lock xlator handles lk_owners
10:24 pranithk soumya_: why does syncopctx_setfslkowner assume the void* is gf_lkowner_t*?
10:24 soumya_ pranithk, this was the same thing which I was trying to talk to you last week ..
10:24 pranithk soumya_: you asked me why lk-owner is used in gluster. But I still lack the context about this patch.
10:24 ndevos soumya_: sure, this would only be one piece - ganesha still would need to use it and all
10:25 ndevos soumya_: why would ganesha need access to a lk-owner, when samba can do without?
10:25 soumya_ ndevos, I still doubt it..because ganesha has support to handle the locks even if the FSAL cannot distinguish locks set by different NFS clients
10:26 soumya_ ndevos, this patch is not at all required for Ganesha..
10:27 soumya_ Ganesha can handle locks/delegations irrespective of whether underlying FSAL can/not differentiate locks..
10:27 ndevos soumya_: ah, right, but why have lk-owner in a public API? that should really only be internal
10:27 soumya_ on the other hand, SAMBA seem to be having issue with locking...
10:27 soumya_ poornima had recently checked it out while I was discussing about lk_owners with her..
10:29 ndevos how do we keep track of lk-owners on the client side? is that done in the rpc-layer?
10:30 soumya_ ndevos, agree..'lk_owner' shouldnt be in public API, moved it to 'void *'...I had added that API to have a API similar to "glfs_setfsuid/gid/groups"
10:30 soumya_ but as I had mentioned in the comment, we can chose not to have that API
10:30 soumya_ and let lock routines handle / fill the field if required.
10:31 ndevos yeah, I do not think we should need the API bit
10:32 * ndevos pokes his browser again, and hopes Gerrit comes to life
10:32 soumya_ ndevos, sound fine for me too.. :)
10:32 soumya_ ndevos, I shall revert back the gfapi changes then..
10:33 ndevos soumya_: yes please - and make the subject of the comment a one-liner :)
10:33 soumya_ ndevos, yupp.. thanks :)
10:35 ndevos soumya_: other than that, it makes sense, and I wonder if locking in Samba improves with this
10:35 soumya_ ndevos, it may/may not..depends on how they like to address this issue
10:35 * ndevos wonders why review.gluster.org is the only site that seems to be *extremely* slow
10:36 soumya_ iit atleast gives an option to ditinguish locks
10:36 ndevos yeah, and thats important
10:37 ndevos soumya_: how are lk-owners tracked client side? are they part of an fd_t or something?
10:38 soumya_ FWIU, they are used differently by different clients..
10:38 soumya_ FUSE entirely seems to be depending on what kernel passes it to it..
10:38 ndevos hmm, thats pretty ugly
10:39 ndevos can we get our locks maintainer to fix that?
10:39 soumya_ and am not sure if kernel ties those lk_owners to fds or the  pids
10:40 soumya_ and Gluster-NFS seems to be using nlm4_netobject as lk_owner
10:41 ndevos the kernel would tie it to the pid+uid, I think
10:42 ndevos hmm, sad that there is no real lk-owner structure used then :-/
10:42 soumya_ ndevos, maybe and there is a discussion about having open file descriptor locks support too right ?
10:43 soumya_ ndevos, hmm yes..and these are cases of fcntl locks..am not sure how this field is populated incase of internal locks
10:45 ndevos fuse seems to pass a uint64_t as lk-owner
10:48 ndevos ah, fuse takes the pointer to the files_struct and "scrambles" it - it does not have any fields in it
10:48 anoopcs joined #gluster-dev
10:50 ndarshan joined #gluster-dev
10:50 ndevos in case someone cares: http://git.kernel.org/cgit/linux/kernel/git/​next/linux-next.git/tree/fs/fuse/file.c#n310
10:53 sachin_ joined #gluster-dev
10:54 spandit joined #gluster-dev
11:03 shubhendu joined #gluster-dev
11:05 ndevos REMINDER: in Gluster Community Bug Triage Meeting will happen in ~1 hour in #gluster-meeting
11:06 ira joined #gluster-dev
11:07 ndarshan joined #gluster-dev
11:07 bala joined #gluster-dev
11:10 rjoseph|afk joined #gluster-dev
11:11 nishanth joined #gluster-dev
11:15 soumya_ ndevos, thanks for the link
11:19 hagarth joined #gluster-dev
11:27 ndarshan joined #gluster-dev
11:32 kkeithley1 joined #gluster-dev
11:38 Debloper joined #gluster-dev
11:42 soumya joined #gluster-dev
11:59 hagarth joined #gluster-dev
12:00 soumya_ joined #gluster-dev
12:07 shubhendu joined #gluster-dev
12:17 ndarshan joined #gluster-dev
12:26 kanagaraj joined #gluster-dev
12:42 _Bryan_ joined #gluster-dev
12:42 kkeithley_ does someone want to send a 3.6.2 announce email on raghu's behalf?
13:01 nkhare joined #gluster-dev
13:02 vikumar joined #gluster-dev
13:04 badone joined #gluster-dev
13:05 kkeithley_ is there a gluster-announce@gluster.org? Release checklist says so, but mail to it just bounced.
13:05 rtalur_ joined #gluster-dev
13:09 kkeithley_ uh, where's the _new_ blog?   wordpress.gluster.org is the old blob, right? And the front page there is all about 3.5. Maybe it should redirect to the new blog!!!
13:10 kkeithley_ wordpress.gluster.org = blog.gluster.org, fwiw
13:11 rjoseph|afk joined #gluster-dev
13:18 kkeithley_ GlusterFS 3.6.2 GA is released.  See http://blog.gluster.org/2015/0​1/glusterfs-3-6-2-ga-released/ for details
13:31 ndevos kkeithley++ thanks!
13:31 glusterbot ndevos: kkeithley's karma is now 49
13:40 hagarth kkeithley++
13:40 glusterbot hagarth: kkeithley's karma is now 50
13:54 lalatenduM kkeithley, are you around?
13:57 pp joined #gluster-dev
13:58 kkeithley_ yup
13:59 lalatenduM kkeithley, when I pm to you I am getting "[kkeithley] is away (I'm not here right now)"
14:00 lalatenduM kkeithley, both in freenode and in rh irc
14:00 kkeithley_ that's because I'm kkeithley_ when I WFH
14:00 lalatenduM kkeithley_, ahh got it
14:00 kkeithley_ kkeithley is on my desktop at work
14:00 lalatenduM kkeithley_, cool
14:01 kkeithley_ s/work/at the office
14:01 kkeithley_ he's not there right now. ;-)
14:01 lalatenduM kkeithley_, :)
14:05 shaunm joined #gluster-dev
14:12 shyam joined #gluster-dev
14:45 shubhendu joined #gluster-dev
15:07 jobewan joined #gluster-dev
15:07 kkeithley_ GlusterFS Office Hours at FOSDEM, details at http://blog.gluster.org/2015/01/g​lusterfs-office-hours-at-fosdem/
15:18 lpabon joined #gluster-dev
15:19 nkhare joined #gluster-dev
15:26 wushudoin joined #gluster-dev
15:28 _Bryan_ joined #gluster-dev
16:14 kkeithley1 joined #gluster-dev
16:15 aravindavk joined #gluster-dev
16:15 jamesc|2 joined #gluster-dev
16:57 _Bryan_ joined #gluster-dev
17:13 shubhendu joined #gluster-dev
17:29 kanagaraj joined #gluster-dev
17:33 gem joined #gluster-dev
18:10 lalatenduM joined #gluster-dev
18:23 shubhendu joined #gluster-dev
19:18 shubhendu joined #gluster-dev
19:33 tdasilva joined #gluster-dev
19:45 ilbot3 joined #gluster-dev
19:45 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/
20:58 ira joined #gluster-dev
21:21 jriano joined #gluster-dev
22:38 bala joined #gluster-dev
23:15 bala joined #gluster-dev

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