Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2014-11-05

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

All times shown according to UTC.

Time Nick Message
00:19 badone_ joined #gluster-dev
01:47 ira joined #gluster-dev
02:06 bala joined #gluster-dev
02:21 ira joined #gluster-dev
02:49 hchiramm joined #gluster-dev
02:49 bharata-rao joined #gluster-dev
03:23 ira joined #gluster-dev
03:34 shubhendu__ joined #gluster-dev
03:35 bharata-rao joined #gluster-dev
03:45 kshlm joined #gluster-dev
04:10 nishanth joined #gluster-dev
04:30 anoopcs joined #gluster-dev
04:37 ira joined #gluster-dev
04:40 rafi1 joined #gluster-dev
04:40 Rafi_kc joined #gluster-dev
04:46 atinmu joined #gluster-dev
04:46 lalatenduM joined #gluster-dev
04:52 spandit joined #gluster-dev
04:58 badone_ joined #gluster-dev
05:01 kanagaraj joined #gluster-dev
05:03 soumya joined #gluster-dev
05:11 ndarshan joined #gluster-dev
05:11 aravindavk joined #gluster-dev
05:14 ira joined #gluster-dev
05:20 hagarth joined #gluster-dev
05:25 nkhare joined #gluster-dev
05:28 atalur joined #gluster-dev
05:39 Humble ndevos, lalatenduM kkeithley 3.5.3beta2 is ready at download.gluter.org http://download.gluster.org/pub/glust​er/glusterfs/qa-releases/3.5.3beta2/
05:41 Humble please cross check and announce ..
05:46 lalatenduM Humble, checking
06:00 kshlm joined #gluster-dev
06:00 kdhananjay joined #gluster-dev
06:15 bala joined #gluster-dev
06:15 anoopcs1 joined #gluster-dev
06:19 ppai joined #gluster-dev
06:20 anoopcs joined #gluster-dev
06:25 anoopcs joined #gluster-dev
06:33 kaushal_ joined #gluster-dev
06:37 kshlm joined #gluster-dev
06:38 badone_ joined #gluster-dev
06:39 raghu` joined #gluster-dev
06:43 lalatenduM Humble, looks fine Humble++
06:43 glusterbot lalatenduM: Humble's karma is now 10
06:50 nishanth joined #gluster-dev
06:53 azar joined #gluster-dev
07:04 Humble lalatenduM++ u too!!
07:04 glusterbot Humble: lalatenduM's karma is now 36
07:04 lalatenduM Humble, :)
07:06 topshare joined #gluster-dev
07:07 Humble I will wait for next scratch builds :)
07:19 aravindavk joined #gluster-dev
07:19 atinmu joined #gluster-dev
07:26 jiffin joined #gluster-dev
07:28 ppai joined #gluster-dev
07:28 Humble JustinClift++
07:28 glusterbot Humble: JustinClift's karma is now 28
07:37 nishanth joined #gluster-dev
07:38 rgustafs joined #gluster-dev
07:58 ppai joined #gluster-dev
08:18 xavih JustinClift: yes, we have already seen it. I think I forgot to answer an email to davemc when he requested me. We discussed the text to send, but it seems I never sent it :-/. My fault...
08:18 xavih JustinClift: thanks for pointing it out
08:32 ira joined #gluster-dev
08:43 vimal joined #gluster-dev
08:43 ndevos Humble++ lalatenduM++ thanks guys! I'll blog it now
08:43 glusterbot ndevos: Humble's karma is now 11
08:43 glusterbot ndevos: lalatenduM's karma is now 37
08:44 Humble ndevos, np!!
08:47 ndevos lalatenduM, Humble: http://blog.nixpanic.net/2014/11/gl​usterfs-353beta2-release-notes.html will make it into blog.gluster.org in a bit
08:48 Humble ndevos, please mention to reach out to
08:48 Humble gluster mailing lists
08:48 Humble and also point 3.5.3 tracker bugs
08:48 ndevos Humble: working on the email already :)
08:49 Humble u can mention in the blog as well :)
08:49 Humble otherwise it looks good to me
08:49 Humble ndevos++
08:49 glusterbot Humble: ndevos's karma is now 44
08:50 bala joined #gluster-dev
08:51 shubhendu__ joined #gluster-dev
08:52 nishanth joined #gluster-dev
08:52 ndevos Humble: yeah, when the email is in the archives, I'll add a link to it in the blog post
08:52 ndevos and in all the fixed bugs
08:52 Humble ok.. cool
09:06 ndevos Humble: email sent, blog has a link now too - going to update the bugs to ON_QA
09:06 Humble ndevos, checking
09:06 ira joined #gluster-dev
09:25 Humble ndevos LGTM
09:25 ndevos Humble: will you blog about the job too? have it included in Fedora Planet?
09:26 Humble hmmmmm.. let me think
09:27 ndevos Humble: not sure if you have your blog setup to get synced to planet.fedoraproject.org, see http://fedoraproject.org/wiki/Planet_HowTo
09:27 * ndevos his blog posts get syndicated when he tags it with "Fedora"
09:27 Humble I havnt .. but can do
09:27 Humble thanks for the tip
09:28 ndevos lalatenduM: maybe for you too? ^
09:28 Humble ndevos++ thanks for the reward too :)
09:28 glusterbot Humble: ndevos's karma is now 45
09:28 ndevos Humble: np!
09:29 shubhendu__ joined #gluster-dev
09:33 nishanth joined #gluster-dev
09:48 shubhendu__ joined #gluster-dev
09:50 soumya joined #gluster-dev
10:09 lalatenduM ndevos, will do , thanks for the tip
10:24 nishanth joined #gluster-dev
11:03 aravindavk joined #gluster-dev
11:11 bala joined #gluster-dev
11:23 krishnan_p joined #gluster-dev
11:24 shubhendu joined #gluster-dev
11:26 ira joined #gluster-dev
11:39 ppai joined #gluster-dev
12:02 jdarcy joined #gluster-dev
12:05 edward1 joined #gluster-dev
12:27 bala joined #gluster-dev
12:28 kanagaraj joined #gluster-dev
12:41 kshlm joined #gluster-dev
13:04 kdhananjay joined #gluster-dev
13:07 ndevos kkeithley: ah, I was thinking "I don't see how to do that" was a response on the libgfapi-versioning backport, that made me wonder
13:08 kkeithley nobody answered my question though about backports. 3.5 and 3.4 also?
13:08 ndevos well, I was interested in that too
13:08 ndevos I think we should backport it, yes
13:09 hagarth kkeithley: +1
13:09 ndevos but should we do that now, and make a beta+=1 for next week, or wait for the next release
13:09 kkeithley I don't find an open BZ. Anders' patch was just an RFC too. Guess I'll just open new BZs
13:09 Humble +1 for the backport
13:09 * ndevos isnt aware of a bz for it
13:10 * Humble neither me
13:10 kkeithley we don't have an api or gfapi component?
13:11 ndevos maybe libgfapi?
13:11 kkeithley yeah, two seconds faster than me
13:11 * ndevos only guesses
13:13 Humble yep.. libgfapi
13:13 Humble libgfapi and libgfapi-python are present
13:18 kaushal_ joined #gluster-dev
13:18 kkeithley https://bugzilla.redhat.co​m/show_bug.cgi?id=1160709
13:18 glusterbot Bug 1160709: unspecified, unspecified, future, kkeithle, POST , libgfapi: use versioned symbols in libgfapi.so for compatibility
13:19 topshare joined #gluster-dev
13:19 * ndevos goes for lunch
13:30 kkeithley ndevos: when you get back from lunch——   you don't have a tracker yet for 3.5.4, or do you want to squeeze this into 3.5.3?
13:31 topshare joined #gluster-dev
13:33 JustinClift ndevos: So keybase.io seems useful... but it also seems like they've set out to people not trust them at all
13:34 JustinClift ndevos: They're a *US* company... and they're asking people to upload their GPG/PGP _private_ keys as well as the rest.
13:34 JustinClift It's optional, but jeeeez.
13:34 kkeithley pfft
13:34 kkeithley good luck with that
13:34 JustinClift Exactly
13:34 JustinClift Just asking for that raises a _lot_ of questions about their intentions
13:35 JustinClift Obviously maximal user security isn't on the list
13:36 kkeithley Not sure I'm so worried about their intentions. Just the fact that they have something that can be obtained with a supoena
13:42 JustinClift Yeah.  But the two are linked.  They'd have to be aware they have to hand over the data when asked, and yet they still do.
13:42 JustinClift Anyway... back to work
13:42 JustinClift Actually, time for coffee and to head into the Baker St office
13:43 JustinClift Cards arrived already. :)
13:46 kkeithley oh good
13:48 nkhare joined #gluster-dev
13:51 topshare Have anyone use zfs in Gluster?
13:52 topshare Or use SSD Cache way to impove the vloume performance?
13:55 rafi1 joined #gluster-dev
13:57 ira joined #gluster-dev
13:58 kkeithley there's this http://www.gluster.org/community/do​cumentation/index.php/GlusterOnZFS
13:58 kkeithley for a start
13:59 ndevos you can also use dm-cache for putting an SSD in front of your LV
14:01 bala joined #gluster-dev
14:06 shyam joined #gluster-dev
14:07 kkeithley oh merde.  "private" functions glfs_new_from_ctx() and glfs_free_from_ctx() were added after 3.6 branched. Anders had suggested that GFAPI_PRIVATE symbols carry a version too. I hadn't checked, presumed all were there since 3.4.0
14:10 ndevos more fun!
14:19 kkeithley I can change/add GFAPI_PRIVATE to be GFAPI_PRIVATE_x.x.x. On master glfs_{new,free}_from_ctx() can be GFAPI_PRIVATE_3.7.0. If we backport those to, e.g. 3.6.2, then they'll have to be changed accordingly on master. That shouldn't be an issue. We don't ship the master branch and it's understood to be a WIP.
14:19 * kkeithley wanted to be done with this
14:21 ndevos yeah, I think we should add additional versions while doing the backport, not change the version in the master branch
14:21 ndevos hmm, maybe we need to add the version of the backport to the master branch too..
14:23 kkeithley I don't follow you.   glfs_{new,free}_from_ctx() are only in master.  Ostensibly they need some version that's not GFAPI_PRIVATE_3.4.0.
14:23 ndevos oh, if they are only in master, thats fine
14:23 kkeithley and if (or when) they are backported to 3.6.2+
14:23 kkeithley ?
14:24 ndevos I thought they got backported, but that may well be only in RHS (and we should not need to care about that)
14:24 kkeithley then they are added as GFAPI_PRIVATE_3.6.2 on release-3.6 branch, and we change the version on master?
14:24 kkeithley but RHS doesn't have versioned symbols yet (if ever)
14:25 kkeithley maybe they never get backported and remain at GFAPI_PRIVATE_3.7.0
14:25 ndevos RHS is a non issue for these symbols in general, glusterfs-api and glusterfs-server will always have the same version, and these symbols are only used by gluster/nfs
14:26 ndevos 3.7.0 should be good, when we backport a symbol, we should add the backport-version to the list (public or private)
14:27 ndevos by adding a version and not removing one, we wont break anything that others are doing (thinking about things like samba/ganesha developers using master)
14:30 kkeithley but changing a version.  if glfs_{new,free}_from_ctx() change from GFAPI_PRIVATE_3.7.0 to GFAPI_PRIVATE_3.6.x at some point — some point before we branch release-3.7 — then any pain as a result is just developer pain.
14:30 JustinClift Whoo that's overnight emails processed. Time to head into office. :)
14:32 ndevos developer pain is still a pain, and it may not be obvious what is wrong when the versions dont match
14:32 sickness oh 3.7 already on the horizon? nice =_)
14:34 kkeithley 3.7 in April 2015
14:35 ndevos kkeithley: whats the change in patch #17 ?
14:35 kkeithley yes, pain is pain. That's why I'm asking the question now. But I don't see a good solution.
14:35 kkeithley Dunno. I edited the commit message in gerrit. It seems to have triggered the change all by itself
14:35 ndevos what is the issue with providing two versions for a symbol - only for the symbols that get backported?
14:36 ndevos oh, okay, yes, a change in the commit message does that
14:36 ndevos we should try not to backport symbols, that aside
14:37 rgustafs joined #gluster-dev
14:39 kkeithley ????  Sure. But, e.g. we took new bits in 3.4.2 so that we could have the gluster FSAL in ganesha earlier than if we had waited for 3.5. And again for 3.5.1 for something ganesha related.
14:39 kkeithley I'd be happy if we could commit to never backporting glfs_{new,free}_from_ctx() to release-3.6.
14:39 itisravi joined #gluster-dev
14:40 ndevos oh, I agree, that definitely should not be needed
14:40 kkeithley I guess we'll presume that we won't backport glfs_{new,free}_from_ctx(). If that changes then we'll cross that bridge when we get to it
14:41 ndevos it should not get backported, we should fix it an provide syncop_* replacement functions instead
14:41 ndevos iff that functionality should get included in release-3.6
14:42 kkeithley sounds like a plan
14:42 ndevos well, we should do the syncop_*() replacement never the less
14:44 ndevos oh, yet an other one hitting installation problems with 3.6 on CentOS :-/
14:46 ndevos kkeithley: can you fix the branch/topic of the patch too? you can use the Gerrit webui and set it to bug-1160709
14:50 kkeithley sure. I'm about about to submit another patch anyway. Perhaps that will that fix it automatically?
14:51 kkeithley ugh, maybe I need to clone a new tree and fetch patch set 17
14:51 ndevos maybe, yes, sometimes Gerrit fixes the rfc -> bug-N automatically, sometimes not - I dont get it
15:12 wushudoin joined #gluster-dev
15:24 kkeithley ndevos: what was your answer to squeezing symbol versions into 3.5.3? Since there's no tracker BZ yet for 3.5.4
15:24 kkeithley Or will you open a tracker BZ for 3.5.4
15:35 ndevos kkeithley: do you want to include the versions in 3.4.6?
15:35 * ndevos thinks it is low-risc and would include it in 3.5.3 if you'd like that
15:41 ndevos how many revisions of a patch should someone post before it gets merged automatically?
15:47 lalatenduM ndevos, kkeithley suggestion, should we wait and see how symbol versions doing in 3.6 then merging them in 3.4 and 3.5?
15:48 * lalatenduM thinks how to minimize any disruption if it comes
15:53 ndevos lalatenduM: its an option, I do not have a strong preference for either way
15:54 lalatenduM ndevos, ok
15:58 ndevos lalatenduM: what is your preference on installing 3.4 or 3.5 on rhel/centos-6.6? use exclude/includepkgs, or yum-plugin-priorities?
15:59 ndevos Humble, kkeithley: your opinion is appreciated too :)
16:00 lalatenduM ndevos, yes, and during update there is another option as "yum-versionlock"
16:00 lalatenduM ndevos, I mean exclude/includepkgs, or yum-plugin-priorities as choice
16:00 kkeithley I'd say that with lots of revisions you should be less likely to automatically merge
16:01 lalatenduM let user decide what method he want to use
16:01 ndevos lalatenduM: I dont think versionlock is very appropriate
16:02 ndevos kkeithley: indeed, when a patch hits #20, the submitter should get a warning?
16:02 kkeithley that's why I stopped at #19
16:02 lalatenduM ndevos, I am thinking for users updating 6.5 to 6.6, it is another option , may be not be a good one as you said
16:02 ndevos yes, you're close!
16:03 ndevos lalatenduM: nah, that does not sound like good advise
16:03 ndevos it will work, but well...
16:04 ndevos lalatenduM: so, I'll document both approaches, test them and blog/email/wiki that?
16:08 lalatenduM ndevos, regarding yum versionlock , I also not sure , just wanted to give all the possible alternative , may be I will put a comment in blog abt it :)
16:08 ndevos lalatenduM: sure, works for me :)
16:09 lalatenduM ndevos, yes, blog would be fine , and same contents on wiki and email . it this getting toomuch or boring for you, let me knwo how I can help
16:09 lalatenduM :)
16:09 lalatenduM s/it is /if it is/
16:12 ndevos lalatenduM: nah, its fine, I should have things ready soon
16:13 ndevos lalatenduM: I think we should build a gluster-release package for the different versions, so that the .repo file with priority=50 can get included there
16:13 ndevos lalatenduM: maybe that is something you would like to take on?
16:13 ndevos lalatenduM: https://github.com/kalebskeithley/gluster-release would be a start for that
16:17 lalatenduM ndevos, yes I can take that. But I am still not comfortable with priority in it
16:18 ndevos lalatenduM: how about commenting out the priority and adding a link to the wiki?
16:18 kkeithley seems to me like a Real RHEL box can be subscribed to channels but there's no /etc/yum.repos.d/*.repo in which you could add a line like "exclude=glusterfs*"
16:18 ndevos kkeithley: add it to /etc/yum.conf
16:18 ndevos or, drop a file in /etc/yum.conf.d/
16:19 kkeithley what will that do if I add my own repo file for CentOS Storage SIG or d.g.o repo?
16:19 ndevos that additional repo should have "includepkgs=glusterfs*"
16:20 ndevos exclude by default, only include from selected sources
16:20 lalatenduM ndevos, yeah thats fine :)
16:20 kkeithley the implication being that the includepkgs= line over-rides the exclude= line
16:20 ndevos yes, that is how it works
16:21 kkeithley okay, that answers that. ;-)
16:22 ndevos it is easier than editing all .repo files (like centos.repo, centos-updates.repo, ... - if those exist)
16:25 kkeithley <non sequitur>I created a SLES12 vm to do builds on. twice I've requested an eval license like I did for SLES11sp3, so I can add the extra -devel packages need to build. No response.</non sequitur>
16:25 kkeithley No SLES12 packages for you SuSE
16:26 kkeithley I should tweet that to #suse and see if I get a better response
16:36 jiffin joined #gluster-dev
16:41 jiffin joined #gluster-dev
16:43 nishanth joined #gluster-dev
16:45 hagarth joined #gluster-dev
16:47 jiffin joined #gluster-dev
16:53 jobewan joined #gluster-dev
16:53 jiffin joined #gluster-dev
16:59 ndevos dammit, the exclude=glusterfs* in the yum.conf disables the includepkgs= in the .repo files - it works fine on fedora :-/
17:20 ndk joined #gluster-dev
17:22 portante joined #gluster-dev
17:26 edward1 joined #gluster-dev
17:32 jiffin joined #gluster-dev
17:33 kkeithley ready for patch #20?
17:35 jiffin joined #gluster-dev
17:36 * ndevos ducks
17:36 ndevos and, yum-plugin-priorities is not in standard rhel
17:37 ndevos blocking the rhel version of glusterfs seems to have quite some edges
17:38 jiffin joined #gluster-dev
17:39 ndevos but, it is in the optional channels... at least that
17:40 lkoranda joined #gluster-dev
17:40 ndevos kkeithley: whats the change for #20?
17:41 kkeithley just version numbers. handle support was backported to 3.4.2 for ganesha
17:41 ndevos oh, ok
17:52 lalatenduM joined #gluster-dev
18:13 jiffin joined #gluster-dev
18:22 jiffin1 joined #gluster-dev
18:37 kkeithley backing up a bit. I don't want to try to squeeze symbol versions int 3.4.6 (or 3.5.3). I was really just poking you to create a 3.5.4 tracker BZ, but giving you the option of squeezing it into 3.5.3 instead
18:37 kkeithley s/int 3.4.6/into 3.4.6/
18:40 kkeithley what does this mean?
18:40 kkeithley ! [remote rejected] HEAD -> refs/for/release-3.4/bug-1125245 (change 9015 closed)
18:40 kkeithley error: failed to push some refs to 'ssh://kkeithle@git.gluster.org/glusterfs.git
18:44 kanagaraj joined #gluster-dev
18:59 _Bryan_ joined #gluster-dev
19:16 semiosis kkeithley: most likely the remote is ahead of you.  try a pull/merge then pushing again
19:20 ndevos kkeithley: http://review.gluster.org/9015 has been merged already, so you can not update that Change-Id
19:33 kkeithley yeah, I fat fingered something. I've since redone it and rfc.sh'd it to gerrit
19:38 lalatenduM kkeithley, here is the specfile for 3.4.6 beta2 https://github.com/LalatenduMohanty/Glus​ter-rpm-specs/blob/3.4.6/glusterfs.spec
19:39 lalatenduM kkeithley, and these are the changes beta1 vs beta2 https://github.com/LalatenduMohant​y/Gluster-rpm-specs/commit/23fb7f8​c355d4546202043540bdf91ac08c704a9
20:17 ndevos Humble: http://blog.nixpanic.net/2014/11/insta​lling-glusterfs-34x-35x-or-360-on.html is live, I hope it is clear enough
20:36 barbara joined #gluster-dev
20:42 Guest48097 left #gluster-dev
20:55 shyam joined #gluster-dev
21:13 kkeithley ndevos++
21:13 glusterbot kkeithley: ndevos's karma is now 46
23:10 badone joined #gluster-dev

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