Camelia, the Perl 6 bug

IRC log for #gluster-dev, 2013-06-05

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

All times shown according to UTC.

Time Nick Message
01:15 xavih joined #gluster-dev
01:34 itamar_ joined #gluster-dev
01:42 itamar_ hi,
01:43 itamar_ I have a issue with glusterfs,
01:43 itamar_ php -q x.html PHP Fatal error:  Unknown: Failed opening required 'x.html'
01:43 itamar_ cat x.html works
02:01 avati itamar_: can you get the output of 'strace php -q x.html'?
02:01 avati and check which syscall is failing?
02:04 hagarth johnmark: around?
02:07 hagarth quick status update on 3.4.0 - one more beta to be pushed out this week.
02:07 itamar_ avati, how can I upload the results of strace to fpaste ?
02:07 hagarth avati: regarding api inclusion, can we schedule that for 3.4.1?
02:09 avati hagarth: the reason i brought it up was because the set of changes is largely isolated within api/src/ and does not affect the rest of the system for general use, and it would be nice to get samba support in fedora earlier
02:10 avati it's only a matter of straight forward backporting and packaging to deliver the feature in fedora, with no change to the general data path
02:11 johnmark hagarth, howdy
02:11 hagarth avati: I agree that changes are isolated - given the churn with any patchset, I was wondering if we could push it beyond 3.4.0. The packaging change can happen right away.
02:11 hagarth johnmark: hey there
02:12 johnmark avati, what about the things that rely on gfapi?
02:12 avati the samba team also keeps asking about glusterfs version which has the api support, and there isn't one
02:12 avati johnmark: samba and qemu
02:13 avati johnmark: are the only two
02:13 hagarth avati: we have had this with qemu for a while too.
02:13 johnmark avati, right. are we sure this wont screw up qemu?
02:13 hagarth on an unrelated note, we have had a similar problem with ovirt too.
02:13 johnmark can we backport and test in a single day?
02:13 avati johnmark: fairly confident, samba exercises gfapi alot more heavily than qemu
02:14 johnmark ok
02:14 hagarth johnmark: these changes should not affect qemu.
02:14 johnmark hagarth, when were you thinking for beta 3?
02:14 hagarth johnmark: this week.
02:15 hagarth johnmark: maybe today or tomorrow before i start spending time in the air :).
02:15 johnmark hagarth, i support the backport that it can be merged and pass some default testing by this time tomorrow
02:15 avati johnmark: it will be good to have both samba and qemu "supported" in 3.4.0.. easier for samba guys too to get RPMs and enable it in their build (i have provided custom rpms for them to test, it will be really good if they start using standard release rpms asap)
02:15 johnmark avati, oh i definitely see the benefits
02:15 hagarth avati: would you want rpm separation for libgfapi?
02:16 avati hagarth: that's orthogonal, but i would like that separation too
02:16 johnmark avati, just worried about unintended consequences, which always come up
02:16 johnmark hagarth, makes sense to me
02:16 hagarth avati: let us do it right when we GA libgfapi. so let's do the backporting and separation together.
02:17 hagarth "separation together" -- almost an oxymoron :).
02:17 avati hagarth: then do we even want libgfapi.so as part of glusterfs.rpm in 3.4.0?
02:18 avati since the larger theme of the release itself is around gfapi and its ecosystem, it would be nice to have it supported well in the first release
02:18 JoeJulian unless we're going to split it off into glusterfs-libs
02:19 johnmark joejulian, that might make sense, and then have a libs-devel rpm
02:19 JoeJulian That's sort-of the standard way of doing it.
02:19 avati right.. having libgfapi.so in glusterfs.rpm in 3.4.0 and then in glusterfs-libs.rpm in 3.4.1 would be confusing
02:19 johnmark agreed.
02:19 hagarth avati: right, I think having both backport and separation in 3.4.0 is a good idea.
02:20 itamar_ avati http://fpaste.org/16591/37039879/
02:20 johnmark avati, so we make the -libs rpm for 3.4.0
02:20 JoeJulian +1
02:20 avati i propose glusterfs-api and glusterfs-api-devel, but -libs is fine too
02:21 hagarth itamar_: which version are you using?
02:21 avati -api is more clear (that it is the gfapi package), and -libs could mean libglusterfs.so too
02:21 JoeJulian The api version is usually done as a "Provides"
02:22 hagarth if we make it api-devel, would it be hard to explain the difference between glusterfs-devel and glusterfs-api-devel?
02:22 itamar_ hagarth 3.3.1
02:22 avati what is glusterfs-devel really?
02:22 avati it was for hekafs mostly
02:23 avati iirc
02:23 avati we could have a common glusterfs-devel
02:23 avati which has all headers
02:23 johnmark yeah, -libs is the customary way to do this
02:23 avati the -api is for separately upgrading libfapi
02:23 itamar_ hagarth, in line 11 it reads all contents of the file  and in line 44 you see the php error.
02:24 hagarth avati: that sounds like a better model
02:24 avati JoeJulian: "Provides" is for the version, this is more for the package name to make the meaning clear.. glusterfs-libs woudl typically be libglusterfs.so libgfrpc.so libgfxdr.so
02:24 johnmark avati, ah ok. hrm
02:25 avati JoeJulian: but we have those libs really in core glusterfs.rpm
02:25 hagarth so api contains libgfapi.so and devel contains glfs.h?
02:26 avati so if someone wants to use qemu or samba integration, they isntall qemu.rpm samba.rpm glusterfs-api.rpm (which depends on glusterfs.rpm)
02:26 johnmark avati, can you see gfapi development actualy changing within .z releases of glusterfs?
02:26 avati if someone wants to develop using gfapi, they install glusterfs-devel.rpm (which depends on glusterfs-api.rpm)
02:26 johnmark avati, ok, now i understand
02:27 hagarth johnmark: since libgfapi is nascent, there will be some churn.
02:27 avati johnmark: the goal is that fixing things in gfapi (since it is still early and new) shouldn't involve uprading glusterfs.rpm (which means restarting glusterfsd services etc)
02:27 hagarth itamar_: anything in the glusterfs client log?
02:27 johnmark avati, and the smb and qemu use cases wouldnt requore other gluster bits on the client side?
02:28 JoeJulian Yeah, there's nothing in the Naming Guidelines so either way's good.
02:28 johnmark avati, agreed
02:28 hagarth hmm, should we rename glusterfs-fuse to glusterfs-client?
02:28 johnmark so then -api and -api-devel? or just toss all headers into glisterfs-devel?
02:29 johnmark im fine either way
02:29 avati johnmark: just -api and all headers tossed into -devel
02:29 johnmark ok
02:29 avati no -api-devel
02:29 johnmark ok
02:29 johnmark hagarth, would that confuse anyone?
02:29 avati hagarth: -fuse is probably better for now.. because gfapi is a -client too
02:29 * hagarth checks the contents of glusterfs-fuse
02:30 avati it has only mount/fuse.so
02:30 avati and mount.glusterfs i think
02:30 hagarth avati: right, for a small interval I was thinking that some client xlators are part of that package too.
02:31 JoeJulian mount/umount, logrotate, and fuse.so
02:31 avati hagarth: we used to do that long ago..
02:31 JoeJulian They were going to be, but they ended up being common.
02:32 avati johnmark: btw vfs_glusterfs.c is commited in forge
02:32 hagarth avati: ok, the model sounds good.
02:33 johnmark avati, excellent
02:33 avati johnmark: forge contains the samba-3.6.x backport, upstream samba.git has it in master (for v4.x)
02:33 hagarth avati: it would be good to include these changes in beta3.
02:34 avati hagarth: i'll send the gfapi backports right away
02:34 johnmark avati, are those in glusterfs-core repo or gluster-samba?
02:34 avati johnmark: samba-glusterfs-vfs.git
02:34 johnmark excellent
02:34 * johnmark looks
02:35 avati https://forge.gluster.org/samba​-glusterfs/samba-glusterfs-vfs
02:35 hagarth johnmark: any progress on moving docs to forge?
02:35 johnmark perfect - thanks
02:35 johnmark hagarth, no.
02:35 johnmark its something ill take care of before summit
02:35 hagarth johnmark: when can we have that populated?
02:36 johnmark hagarth, i will look to do it tomorrow to coincide with beta3
02:36 hagarth johnmark: ah ok, once that is ready let us open that up in gerrit as well.
02:36 johnmark ok
02:36 hagarth johnmark: great!
02:36 hagarth beta3 looks exciting now!
02:37 johnmark hagarth, agreed about including gfapi changes in beta3
02:37 * johnmark crosses fingers
02:37 hagarth maybe it is a good time to get the other jenkins server functional too
02:38 hagarth so that qemu-libgfapi changes also start getting tested.
02:38 johnmark hagarth, any other patches of concern for beta3?
02:38 johnmark other backports in the queue?
02:38 hagarth johnmark: no, op-version will be in today. Once avati's changes are in, it is going to be a period of bug fixes only.
02:39 * johnmark does the happy dance
02:39 johnmark ok then
02:39 hagarth can glusterbot record meetings too?
02:40 itamar_ hagarth, for me nothing in the logs, where I should look ?
02:40 JoeJulian Yes. I'll get that set up.
02:40 hagarth JoeJulian: great, thanks!
02:40 johnmark joejulian, cool
02:41 hagarth itamar_: nothing in /var/log/glusterfs/<mnt-point>.log?
02:41 hagarth johnmark: sounds like we are inching closer - let us keep the momentum going.
02:41 johnmark zodbot in #gluster-meeting can do that, but we dont really use that chanmel for these meetings
02:41 johnmark coolio
02:42 johnmark hagarth, aight. omward and upward. will let you know when the docs deed is done
02:42 johnmark later
02:42 hagarth johnmark: sounds good, later!
02:42 johnmark :-)
02:46 itamar_ hagarth, nothing in the logs
02:48 itamar_ hagarth: background  meta-data data self-heal failed on /dados/scripts/dns/adddns
02:50 hagarth itamar_: is the file in split-brain between the two replicas?
02:51 itamar_ yes
02:52 hagarth itamar_: do you know which copy is the preferred one?
02:53 itamar_ I have created this file for testing
02:53 itamar_ lets assume that server is is the preffered copy
02:54 itamar_ hagarth, lets assume that the server 1 is the preffered copy
02:54 hagarth itamar_: can you delete a copy from one of the bricks?
02:54 hagarth itamar_: in that case, can you delete it from server2?
02:55 itamar_ hagarth, it will probably delete it from both machines.
02:55 hagarth itamar_: from the brick directly .. not from the mount
02:55 itamar_ ok, hold
02:58 shubhendu joined #gluster-dev
02:58 itamar_ hagarth, deleted and now ?
03:02 itamar_ hagarth, I still see the file on server 1
03:03 itamar_ failed to open as split brain seen, returning EIO
03:03 hagarth itamar_: now doe stat on the mount point succeed?
03:04 itamar_ same error in server 1
03:05 hagarth itamar_: how many files do you have in /dados/scripts/dns/adddns ?
03:05 itamar_ 4 files
03:07 hagarth itamar_: you can use this procedure for resolving the split-brain - http://joejulian.name/blog/fixin​g-split-brain-with-glusterfs-33/
03:07 itamar_ hagarth, ok , I will try
03:18 mohankumar__ joined #gluster-dev
03:35 mjrosenb joined #gluster-dev
04:28 anands joined #gluster-dev
05:26 hagarth joined #gluster-dev
05:28 lalatenduM joined #gluster-dev
05:29 lala_ joined #gluster-dev
05:53 bala joined #gluster-dev
06:12 raghu joined #gluster-dev
06:18 bulde joined #gluster-dev
06:34 aravindavk joined #gluster-dev
06:46 bulde joined #gluster-dev
07:15 vshankar joined #gluster-dev
09:14 yliu joined #gluster-dev
09:30 raghaven1rabhat joined #gluster-dev
09:30 awheeler joined #gluster-dev
09:30 avati joined #gluster-dev
09:34 blues-man joined #gluster-dev
10:18 badone_ joined #gluster-dev
10:33 badone_ joined #gluster-dev
10:33 avati_ joined #gluster-dev
10:42 Yuan joined #gluster-dev
10:45 raghaven1rabhat joined #gluster-dev
10:45 awheeler joined #gluster-dev
10:47 edward1 joined #gluster-dev
11:03 lpabon joined #gluster-dev
12:03 hagarth joined #gluster-dev
12:30 hagarth joined #gluster-dev
12:34 raghaven1rabhat joined #gluster-dev
12:34 awheeler joined #gluster-dev
13:18 lpabon joined #gluster-dev
13:24 bfoster_ joined #gluster-dev
13:24 ndevos_ joined #gluster-dev
13:34 itamar_ joined #gluster-dev
13:37 kkeithley joined #gluster-dev
13:45 puebele joined #gluster-dev
13:48 bfoster_ joined #gluster-dev
13:48 ndevos_ joined #gluster-dev
13:48 puebele joined #gluster-dev
13:52 badone_ joined #gluster-dev
13:59 portante joined #gluster-dev
14:00 mohankumar__ joined #gluster-dev
14:01 wushudoin| joined #gluster-dev
14:02 sghosh joined #gluster-dev
14:08 lalatenduM joined #gluster-dev
14:16 portante joined #gluster-dev
14:17 lpabon joined #gluster-dev
14:26 avati joined #gluster-dev
14:29 mohankumar joined #gluster-dev
14:29 hagarth1 joined #gluster-dev
14:29 edward2 joined #gluster-dev
14:48 jbrooks joined #gluster-dev
14:48 edward2 joined #gluster-dev
14:48 mohankumar joined #gluster-dev
14:48 puebele joined #gluster-dev
14:48 raghaven1rabhat joined #gluster-dev
14:48 awheeler joined #gluster-dev
14:50 mohankumar joined #gluster-dev
15:06 vshankar joined #gluster-dev
15:43 mohankumar__ joined #gluster-dev
15:50 mohankumar__ joined #gluster-dev
15:53 aravindavk joined #gluster-dev
15:53 portante_ joined #gluster-dev
15:53 kkeithley joined #gluster-dev
15:54 portante_ joined #gluster-dev
15:54 kkeithley joined #gluster-dev
15:56 mohankumar__ joined #gluster-dev
16:04 blues-man joined #gluster-dev
16:13 itamar_ joined #gluster-dev
16:13 Technicool joined #gluster-dev
16:25 jclift_ joined #gluster-dev
17:08 blues-man joined #gluster-dev
17:08 kkeithley joined #gluster-dev
17:08 hagarth1 joined #gluster-dev
17:08 wushudoin| joined #gluster-dev
17:08 foster joined #gluster-dev
17:08 JoeJulian joined #gluster-dev
17:08 semiosis joined #gluster-dev
17:08 hagarth__ joined #gluster-dev
17:12 bulde joined #gluster-dev
17:21 hagarth kkeithley: checking out build.gluster.org
17:34 aravindavk joined #gluster-dev
17:41 awheeler_ joined #gluster-dev
17:44 hagarth kkeithley, johnmark: avati is looking into b.g.o. I am off now.
18:53 sghosh joined #gluster-dev
19:33 awheeler_ joined #gluster-dev
20:59 a2_ build.gluster.org vm was shutdown (no idea why)
20:59 a2_ restarted the VM, all seems OK now
21:30 portante joined #gluster-dev
21:49 __Bryan__ joined #gluster-dev
22:47 tg2 joined #gluster-dev
23:32 jbrooks joined #gluster-dev

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