Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2016-10-28

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

All times shown according to UTC.

Time Nick Message
00:47 shyam joined #gluster-dev
01:30 hagarth joined #gluster-dev
03:03 magrawal joined #gluster-dev
03:26 mchangir joined #gluster-dev
03:41 nbalacha joined #gluster-dev
03:51 sanoj_ joined #gluster-dev
03:58 hgowtham joined #gluster-dev
04:01 Saravanakmr joined #gluster-dev
04:10 shubhendu joined #gluster-dev
04:20 sankarshan_ joined #gluster-dev
04:23 Muthu joined #gluster-dev
05:00 poornima_ joined #gluster-dev
05:05 ndarshan joined #gluster-dev
05:12 jiffin joined #gluster-dev
05:17 ashiq joined #gluster-dev
05:23 Debloper joined #gluster-dev
05:30 gem joined #gluster-dev
05:30 ppai joined #gluster-dev
05:31 karthik_us joined #gluster-dev
05:39 aravindavk joined #gluster-dev
05:50 kdhananjay joined #gluster-dev
05:51 itisravi joined #gluster-dev
05:57 gem joined #gluster-dev
05:58 asengupt joined #gluster-dev
06:03 Bhaskarakiran joined #gluster-dev
06:11 ankitraj joined #gluster-dev
06:19 hchiramm joined #gluster-dev
06:25 pranithk1 joined #gluster-dev
06:28 riyas joined #gluster-dev
06:29 rafi joined #gluster-dev
06:35 atinm joined #gluster-dev
06:40 purpleidea ppai: are you playing with etcdv2 or etcdv3 (clientv3)
06:40 ppai purpleidea: v2 client
06:41 purpleidea ppai: imo forget about v2. mgmt is using v3, and idk any reason to keep on v2, do you?
06:41 ppai purpleidea: we actually use libkv extensively which internally uses v2 client
06:41 nigelb joined #gluster-dev
06:41 ppai purpleidea: https://github.com/docker/libkv
06:42 purpleidea ppai: i'll look at it...
06:42 ppai purpleidea: they will migrate to v3 some time: https://github.com/docker/libkv/issues/103
06:43 ppai purpleidea: Other than use of libkv, I have no technical reason to use v2 client.
06:43 purpleidea ppai: pretty cool, i have previously not seen this lib.
06:44 Muthu joined #gluster-dev
06:44 purpleidea ppai: i wonder if you'd be able to use this, but still use the complex connection endpoint changing semantics which are in the mgmt code base...
06:45 purpleidea iow the stuff where a client changes endpoints.
06:45 ppai libkv will let you change (or choose) backends, without having to change any client code, if at all we decide to move away from etcd some day.
06:46 ppai the client endpoint change should've worked on creating a new instance of etcdclient (or libkv)
06:46 purpleidea ppai: i like this idea a lot... do you know if libkv can let you still poke at backend specific things?
06:47 ppai purpleidea: It does, not via API though. It's pretty easy to modify code that talks to backend.
06:47 ppai purpleidea: https://github.com/docker/libkv/b​lob/master/store/etcd/etcd.go#L55
06:49 purpleidea ppai: my point was suppose we wanted to do something like this: https://godoc.org/github.com​/coreos/etcd/clientv3#Client.SetEndpoints
06:49 purpleidea through libkv
06:50 purpleidea or any other thing that isn't exposed by libkv, is it doable, or do you have to patch and wrap everything?
06:51 kshlm ndevos, Any idea when 3.9 packages are going to land in the gluster39-test repo?
06:51 kshlm I spent the whole morning today testing 3.8.5 assuming they were 3.9.
06:52 * kshlm should have checked version before starting, but he shouldn't have had to.
06:53 ppai purpleidea: we don't have to change client endpoints, we just create a new libkv instance
06:54 ppai purpleidea: change the "addrs" to point to new endpoint here: https://github.com/docker/li​bkv/blob/master/libkv.go#L29
06:54 purpleidea cool, in any case, i'm going to focus on the other parts, but thanks for pointing me to libkv :)
06:55 ppai purpleidea: okay, I'll get this embed issues sorted out. Do you restart embedded server in your mgmt code ?
06:56 purpleidea ppai: i was watching that issue and trying to remember! i remember stopping it and then starting it again was a problem, but i don't remember the specifics. I think it works now in etcdv3. It's easy to test with mgmt...
06:56 purpleidea let me try...
07:00 msvbhat joined #gluster-dev
07:01 purpleidea ppai: interestingly this does _not_ work. maybe the bug was never fixed... i thought this was something to do with the listener not closing properly or something... hmmm let me see if i can find the issue
07:01 ppai purpleidea: I did bump into the issue you had filed: https://github.com/coreos/etcd/issues/6042
07:03 nishanth joined #gluster-dev
07:05 purpleidea ppai: yes, i remember this one! but i thought it might have been due to something in the listener library... hrmm
07:06 purpleidea can github search for issues that you're watching?
07:07 purpleidea maybe this one? i forget: https://github.com/golang/go/issues/4674
07:09 purpleidea ppai: in any case, hope this all helps. i'm going to sleep if you don't need me for anything
07:10 ppai purpleidea: yeah, get some sleep :) the etcd folks will get back on it.
07:10 purpleidea ppai: cool, i think if you provide a v3 example, it will get more love <3
07:10 Muthu joined #gluster-dev
07:11 ppai purpleidea: v3 works fine, I've tried it. v3 API has a Close() call
07:11 purpleidea ppai: oh and btw, i asked about if they're going to implement the elastic thing in core, and i got a vague answer, which was "sort of" and "it's a secret" -- so abstracting away the backend is always a +ive mood if they're playing open source unfriendly
07:11 mchangir joined #gluster-dev
07:13 ppai purpleidea: Ah I see. The store is completely abstracted by libkv, but the clustering part, not so much.
07:14 purpleidea yeah, exactly, even if libkv abstracts client usage, the tricky code is the server part :P
07:18 purpleidea okay night!
07:19 ppai cya!
07:34 nishanth_afk joined #gluster-dev
07:45 devyani7_ joined #gluster-dev
07:47 Muthu joined #gluster-dev
07:59 riyas joined #gluster-dev
08:00 kshlm aravindavk, pranithk1, Have experimental stuff like jbr and others been disabled in 3.9?
08:02 aravindavk kshlm: pranithk1 I don't think we have option to disable, ndevos mentioned we need to remove code, we don't have ./configure option to disable
08:03 kshlm So does this mean that the experimental stuff is shipping in 3.9? Even if it is being used?
08:05 kshlm configure has a `disable-experimental` flag, that is on be default.
08:05 kshlm This needs to be switched to off for 3.9.
08:06 kshlm I'll open a bug to do this and the op-version change I mentioned on the mail thread.
08:06 kshlm This is a blocker for 3.9.0 release.
08:06 kshlm I'll send out a patch to do this ASAP.
08:08 devyani7_ joined #gluster-dev
08:17 aravindavk kshlm: ok
08:17 kshlm 1389675 is tracking this now.
08:27 mchangir joined #gluster-dev
08:33 spalai joined #gluster-dev
08:45 rastar joined #gluster-dev
08:53 pranithk1 kshlm: Not that I know of
08:54 kshlm pranithk1, aravindavk, http://review.gluster.org/15748 disables experimental stuff.
08:54 kshlm Will you guys track it and ensure it get's merged?
08:55 pranithk1 kshlm: I will take care of it. Thanks!
08:56 kshlm Cool. Now I'm off for lunch.
09:01 rraja joined #gluster-dev
09:07 Debloper joined #gluster-dev
09:15 rafi1 joined #gluster-dev
09:23 devyani7_ joined #gluster-dev
09:37 rastar joined #gluster-dev
09:43 kshlm pranithk1, So the change failed somke because rpmbuild failed. I need help to fix the spec file or I could just delete all experimental code.
09:43 kshlm What's your opinion?
09:44 kshlm Everyone else is welcome to say what I need to do here.
09:49 rafi1 joined #gluster-dev
10:01 ndevos kshlm: I prefer to remove incomplete code, that prevents anyone from trying to build those experimental featues from an old(er) branch
10:01 kshlm ndevos, Then I'll do that.
10:01 rafi joined #gluster-dev
10:01 kshlm I'll anyways port the original patch forward.
10:02 kshlm That shold make it easier to disable experimental if required.
10:02 ndevos yeah, if disabling the experimental builds is not working correctly, we need to fix that in any case
10:06 pranithk1 kshlm: don't delete experimental code.
10:07 kshlm pranithk1, They actually have no use in 3.9
10:07 kshlm They still remain in master.
10:07 pranithk1 kshlm: May be you should send a mail on gluster-devel. Let jeff and shyam respond?
10:08 pranithk1 kshlm: It is better to consult and communicate it before doing that IMO
10:08 kshlm I don't think they'll object.
10:08 kshlm How much longer are you willing to wait for this to get done?
10:08 ndevos for 3.8 both of them agreed that removing was better than disabling, it encourages users to try the experimental features from master and do not expect backports
10:09 pranithk1 kshlm: We are still waiting for acks... so there is time IMO
10:09 kshlm pranithk1, How much time?
10:09 pranithk1 kshlm: How much time do you want?
10:10 pranithk1 kshlm: I will wait...
10:10 kshlm I don't want much.
10:10 kshlm It's just that getting opinions might take a long time.
10:10 kshlm Given its the weekend.
10:10 kshlm And I will not be available for the next 3 days.
10:11 pranithk1 kshlm: Do one thing. You send out that patch also and send a mail
10:11 pranithk1 kshlm: I will watchout for the responses and merge the relevant patch
10:11 kshlm Okay.
10:12 kshlm My first patch is incomplete though. It needs the spec file to be fixed.
10:14 rafi joined #gluster-dev
10:16 kshlm ndevos, Do you know if it is possible to use automake conditionals in '.in' files?
10:17 ndevos kshlm: no, I've never figured out how
10:17 kshlm Maybe kkeithley knows how.
10:19 ndevos kshlm: for most things, an (empty) variable seems to work, otherwise you could maybe use different targets in a Makefile.am with sed/awk on the file
10:19 ndevos kshlm: what is it that you're trying to do?
10:20 kshlm ndevos, What I want to do is use a automake condition to add the experimental files in the spec file.
10:20 kshlm Right now the fdl xlator is added unconditionally.
10:20 kshlm This caused the failure of my earlier patch disabling experimental features.
10:23 ndevos kshlm: having a variable that expands to 0 or 1, and use "%global enable_fdl @@ENABLE_FDL@@" with "%if (enable_fdl)" and the like?
10:24 ndevos kshlm: or maybe just a variable DISABLE_EXPERIMENTAL? that might exist already
10:24 kshlm That's nearly it. Except I'll make it enable_experimental.
10:25 ndevos there are similar variables for it in the top of the .spec.in, for firewalld and suc
10:25 ndevos h
10:26 kshlm I helped do the firewalld stuff. Can't believe I forgot that.
10:28 kshlm Ah, but that didn't involve any automake substituions.
10:38 kshlm pranithk1, New review. Which deletes stuff.
10:38 kshlm https://review.gluster.org/15750
10:39 pranithk1 kshlm: cool
10:39 pranithk1 kshlm: sent the mail?
10:39 kshlm Typing it now.
10:39 pranithk1 kshlm: cool
10:42 prasanth joined #gluster-dev
10:47 Bhaskarakiran joined #gluster-dev
10:47 ankitraj joined #gluster-dev
10:47 kshlm pranithk1, I've sent it now.
10:48 pranithk1 kshlm: thanks a lot for helping with this. I didn't know about this at all...
11:07 kkeithley I would "turn it around".   Add _with_experimental to the spec file, and pass that to configure.  add an AC_ARG_ENABLE stanza to configure.ac, and then build or not build experimental.  E.g. see the cmocka enablement
11:07 kkeithley would that work?
11:07 kkeithley kshlm, ndevos: ^^^
11:08 kshlm kkeithley, That's what I want/
11:09 kshlm I'll see if I can do that later this evening.
11:09 kshlm I want finish testing GlusterD for 3.9 first.
11:10 kkeithley do you only want fdl disabled/excluded? Or all of experimental?
11:10 nishanth joined #gluster-dev
11:12 nbalacha joined #gluster-dev
11:17 gem joined #gluster-dev
11:47 kkeithley can somebody please review http://review.gluster.org/15733, http://review.gluster.org/15737, and http://review.gluster.org/15743
11:49 kkeithley they're small and should be easy to review.  I need to do backports to 3.9, and would be good to have then in the GA. Otherwise I'll have to patch the builds.
11:49 karolyi joined #gluster-dev
11:50 hchiramm ppai, karolyi is looking for u ;:)
11:51 spalai left #gluster-dev
11:52 karolyi ohai
11:52 karolyi I just posted my question in #gluster
11:56 hchiramm karolyi, I see the chat . :)
11:56 ppai thanks hchiramm
11:56 hchiramm yw
12:02 rafi joined #gluster-dev
12:18 pfactum joined #gluster-dev
12:19 rraja joined #gluster-dev
12:22 jiffin joined #gluster-dev
12:30 pranithk1 kkeithley: Could you get reviews? I dont understand them...
12:30 pranithk1 kkeithley: the one where jiffin gave +1 I merged
12:31 kkeithley okay. well, still waiting on the other two, and I'm preparing the release-3.9 versions now
12:32 kkeithley mchangir did give a comment on http://review.gluster.org/15733
12:32 kkeithley mchangir++  did give a comment on http://review.gluster.org/15733
12:32 glusterbot kkeithley: mchangir's karma is now 2
12:34 magrawal_ joined #gluster-dev
12:36 obnox joined #gluster-dev
12:37 ankitraj joined #gluster-dev
12:39 luizcpg joined #gluster-dev
12:47 rafi1 joined #gluster-dev
12:50 luizcpg left #gluster-dev
12:57 poornima_ joined #gluster-dev
12:58 nishanth joined #gluster-dev
13:04 nthomas_ joined #gluster-dev
13:08 shyam joined #gluster-dev
13:16 nishanth joined #gluster-dev
13:23 nishanth joined #gluster-dev
13:44 nishanth joined #gluster-dev
13:47 nbalacha joined #gluster-dev
14:01 nbalacha joined #gluster-dev
14:03 nigelb ndevos: when you're back from vacation, I'd love to clean up the keys on our centos node.
14:04 nigelb It's not urgent, just nice to do.
14:10 nishanth joined #gluster-dev
14:11 lpabon joined #gluster-dev
14:15 karolyi left #gluster-dev
14:16 nishanth joined #gluster-dev
14:18 obnox hagarth: ping
14:18 obnox hagarth: u gonna rejoin?
14:20 ankitraj joined #gluster-dev
14:21 obnox hagarth: dropping off, myself..
14:29 ira joined #gluster-dev
14:34 shaunm joined #gluster-dev
15:05 rraja joined #gluster-dev
15:09 wushudoin joined #gluster-dev
15:39 gem joined #gluster-dev
15:40 luizcpg joined #gluster-dev
15:47 spalai joined #gluster-dev
15:50 ankitraj joined #gluster-dev
16:16 rafi joined #gluster-dev
16:20 magrawal joined #gluster-dev
16:21 rafi joined #gluster-dev
16:42 devyani7 joined #gluster-dev
16:45 mmckeen joined #gluster-dev
16:45 shyam joined #gluster-dev
17:15 hchiramm joined #gluster-dev
17:20 raghu joined #gluster-dev
17:30 luizcpg_ joined #gluster-dev
17:39 dlambrig hi Amye, I want to post a blog to blog.gluster.org. It would be a writeup of some of the Berlin gluster summit work. I used to sign on to gluster.org and be able to edit blogs, then post them. John Mark had given me an account. But I cannnot find where to sign in to do that now and it may have changed. Can you help me?
17:51 kkeithley https://blog.gluster.org/wp-login.php
18:13 dlambrig kkeithley: ++
18:13 glusterbot dlambrig: kkeithley's karma is now 150
19:12 shaunm joined #gluster-dev
21:31 ira joined #gluster-dev

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