Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2017-03-29

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

All times shown according to UTC.

Time Nick Message
00:52 major sooo ... anyone have any input on run-tests-in-vagrant.sh ?
01:12 csaba joined #gluster-dev
01:13 lkoranda joined #gluster-dev
01:34 * major adds ubuntu support...
01:35 major cause like .. why not
01:48 ilbot3 joined #gluster-dev
01:48 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/
04:03 riyas joined #gluster-dev
05:07 ankitr joined #gluster-dev
05:51 msvbhat joined #gluster-dev
06:22 Shu6h3ndu joined #gluster-dev
06:37 Shu6h3ndu joined #gluster-dev
06:39 ankitr joined #gluster-dev
06:50 sanoj joined #gluster-dev
07:10 ankitr joined #gluster-dev
07:36 prasanth joined #gluster-dev
08:35 jiffin joined #gluster-dev
11:06 ndevos nigelb: the CentoOS nightly builds, is that always the epel-6-i386 one that is failing, or have you seen others too?
11:07 ndevos kkeithley: did you not mention something about strfmt issues yesterday? I see failures when building master on fedore-26-i386 :-/
11:09 kkeithley ndevos: I did, but it was an issue with the gnfs-config that went away when I fixed the problem
11:09 ndevos kkeithley: oh, ok
11:09 kkeithley which jenkins job, or which patch?
11:09 ghenry joined #gluster-dev
11:12 ndevos well, I'm looking into doing a GF_ATOMIC_* set of functions, and the rpc bits fail to compile (unmodified)
11:13 ndevos in case you care, http://termbin.com/x17v is the interface I'm trying out
11:14 ndevos the cleanup is minor, but at least everything then has the same behaviour (lock per variable) on all architectures - diffstat http://termbin.com/f7m5
11:19 kkeithley bit where can I see the output from strfmt
11:22 kkeithley s/bit/but/
11:23 * ndevos runs it again
11:24 ndevos that is with "make CFLAGS=-m32" - http://termbin.com/564w
11:25 ndevos its all warnings, and not failures though
11:29 ghenry joined #gluster-dev
11:29 ghenry joined #gluster-dev
12:03 sanoj joined #gluster-dev
12:15 kkeithley ugh, need glibc-devel.i686
12:19 kkeithley or an i686 box
12:23 kkeithley thought I had one
12:28 ira joined #gluster-dev
12:30 kkeithley Perhaps for 4.0 we can drop 32-bit?
12:31 kkeithley vbellur, ndevos: ^^^
12:31 kkeithley shyam: ^^^
12:38 jiffin joined #gluster-dev
12:44 ndevos kkeithley: I doubt it, many distributions will still provide 32-bit versions, and I run 32-bit arm at home
12:44 kkeithley time to upgrade those to aarch64
12:45 ndevos uh, yeah, but I need a nicely boxed versions, currently I have several Wandboards (ethernet + USB-disks)
12:47 ndevos and preferably < 300 Euro :)
12:49 ndevos do we care to build fdl on centos-5-i368? it definitely fails
12:49 ndevos *i386 of course
12:49 kkeithley pine64 and ExpressoBin are < 300€
12:49 kkeithley fdl?
12:49 ndevos full-data-logger, for jbr
12:50 ndevos at least I think it stands for that
12:50 kkeithley RHEL5 finally EOL.  It's too old.
12:50 kkeithley I say drop it
12:52 ndevos are eithere the pine64 or ExpressoBin supported by upsteam kernels?
12:55 kkeithley pbrobinson said the ExpressoBin would work with F25 out of the box. (I should be receiving mine soon.) I have a pair of Pine64 boards, one currently running some debian or Ubuntu respin, but I've heard that F25 now works on it.
12:55 kkeithley Not exactly answering your question
12:55 misc if fedora run, then that's supported upstream
12:58 ndevos yeah, pbrobinson wont include patches that are not upstreamed yet, so that sounds good
13:00 ndevos kkeithley: so, if I get 4 espressobin boards, I can get rid of the armv7hl systems that currently run my storage, we can then drop 32-bit imho :P
13:00 kkeithley it's only money
13:02 kkeithley or you could just keep running 3.10, or 3.11 or 3.12
13:03 kkeithley so, are we having a community meeing today?
13:04 * ndevos does not know
13:04 ndevos got it from https://www.globalscaletechnologies.com/p-72-marvell-espressobin.aspx ?
13:04 kkeithley *meeting*
13:05 ndevos hmm, no case available from there, I would love to see a bunch of them in a rack mount
13:05 kkeithley Mine are coming from their kickstarter
13:05 shyam joined #gluster-dev
13:05 kkeithley kickstarter campaign
13:05 ndevos oh, and a 2GB RAM version is coming soon? I'll wait for that one
13:13 ndevos shyam: there seems to be a libposix2common.so for posix2/dht2? it is placed in a system-libdir, so it would be good to give it a libgf* prefix instead
13:13 msvbhat joined #gluster-dev
13:14 kkeithley ndevos: so, back to your strfmt warnings.
13:15 ndevos kkeithley: I dont really care about them...
13:16 kkeithley gcc -E of rpcsvc.c shows the expansion of the gf_log.  Which uses GF_PRI_RPC_XID, defined in xdr-rpc.h as PRIu32 on DARWIN_HOST and PRIu64 on everything else regardless of 32-bit or 64-bit
13:16 kkeithley oh
13:17 ndevos you want to fix it? I dont know when it was introduced, only noticed them todat
13:17 ndevos *today
13:17 ndevos I think all GF_PRI_* should be in libglusterfs/src/logging.h
13:18 kkeithley Harsha did it, back in 2014
13:18 ndevos oh, maybe I've been just ignoring those errors for a while then
13:18 kkeithley well, he added the DARWIN_HOST PRIu32
13:18 ndevos or warnings
13:18 kkeithley indeed
13:19 ndevos I'm fixing the el5 build, just because I wanted to check on older compilers, and now I hit the most amazing failures
13:19 ndevos the current one is: ../../libglusterfs/src/.libs/libglusterfs.so: undefined reference to `preadv'
13:20 kkeithley RHEL5 doesn't have preadv
13:20 ndevos yeah, I just found out
13:21 kkeithley RHEL5/CentOS5 support died the day we did that.
13:22 ndevos it seems to be, 2015-12-07 was the date it was added to syscall.c
13:22 ndevos and it is only used for storage/bd, that nobody uses
13:23 kkeithley IIRC all those warnings are in trace-level log messages. So no, I really have very little interest in fixing them. (little as in zero)
13:25 kkeithley shyam: ping. where are we writing feature pages for 3.11?
13:25 shyam ndevos: will check and let you know, it is common code for 2 parts of the DHT2 xlator code, so I might have placed it in the system libdir as an shortcut
13:25 shyam kkeithley: gluster-specs and use "Updates: issue#n" in the commit message to reflect that this spec belongs to github issue #n for reference
13:26 kkeithley so I have to open an issue too?
13:26 shyam Feature pages still live in gluster-specs the issue in github helps us tie in everything in (close to) 3 repos in one place, gluster-specs/glusterfs/glusterdocs
13:26 shyam Ummm.. yes it is a feature :)
13:26 shyam No bug though, so replace bug with issue
13:27 kkeithley yeah, sorry, just clarifying that there isn't/wasn't a general issue for 3.11 features
13:27 shyam A general issue for 3.11 feature? As in one issue for all features?
13:28 kkeithley yeah, that's what I was wondering.  Seems the answer is/was no. ;-)
13:28 kkeithley open the issue in gluster-specs issue tracker or gluster issue tracker? Or is there even a gluster-specs issue tracker.
13:29 shyam glusterfs tracker, there is no tracker/issue stuff in gluster-specs repo (I know this is a bit confusing)
13:30 shyam And let me edit my commit message request to state "Updates: gluster/glusterfs#n" (and checking my doc on this for clarity)
13:32 shyam kkeithley: Here are the instructions (see section "This is a very short version (longer version follows)"): https://hackmd.io/GwRgxg7AHCCcYFoIBZkAYHJMgRgghlPhAVGsgGbJjKzIAmATEA==# (which needs to move into dev docs or such, AI on me)
13:39 ankitr joined #gluster-dev
14:05 nigelb kkeithley / ndevos - merged both your changes.
14:05 kkeithley nigelb++ thanks
14:05 glusterbot kkeithley: nigelb's karma is now 48
14:05 nigelb Syncing the change to build nodes now.
14:06 nigelb I didn't want to push a prod change the day before a local holiday :)
14:06 kkeithley isn't today the holiday?
14:07 nigelb It is indeed.
14:07 nigelb Which is why I didn't merge yesterday.
14:07 nigelb If it breaks, I'll see it when I wake up tomorrow :)
14:07 misc we can revert the merge
14:13 nigelb changed synced.
14:13 nigelb break all the builds \o/
14:16 misc perfect, just in time to go home
14:23 kkeithley I kicked off a recheck of CentOS regression of my gnfs-config change and it's running just fine.
14:27 kkeithley netbsd regression seems to be doing fine as well
14:28 kkeithley so we still dont' have an answer on whether we're having a community meeting!
14:29 wushudoin joined #gluster-dev
14:31 wushudoin joined #gluster-dev
14:50 kkeithley I guess we're not having a meeting. Haven't seen any reminders from kshlm
14:53 kshlm|Phone joined #gluster-dev
14:53 kshlm|Phone Sorry for the very last minute notice.
14:54 kshlm|Phone I just realised that the community meeting is today when my phone notified me.
14:55 kshlm|Phone I'm currently out, and will not be able to host the meeting today.
14:56 kshlm|Phone Could someone else could host this meeting, please?
15:01 kkeithley sure
15:01 raghu joined #gluster-dev
15:01 * kkeithley wonders if anyone will attend
15:01 misc there is a company meeting at the same time
15:01 misc and public holiday in India, no ?
15:01 kkeithley that's why I wonder if anyone will attend
15:02 kkeithley I have to find the agenda
15:19 susant joined #gluster-dev
16:05 ankitr joined #gluster-dev
16:53 msvbhat joined #gluster-dev
17:05 raghu joined #gluster-dev
17:25 nishanth joined #gluster-dev
18:07 raghu joined #gluster-dev
18:11 vbellur joined #gluster-dev
18:11 jiffin joined #gluster-dev
18:21 msvbhat joined #gluster-dev
18:27 msvbhat joined #gluster-dev
19:35 raghu joined #gluster-dev
19:51 major so .. is there an active maintainer for the vagrant test infrastructure?
19:55 msvbhat joined #gluster-dev
19:55 major the sound of crickets makes me hungry... https://exoprotein.com/
19:58 major seems like everything needs a code cleanup in order to add new features
19:58 major gonna be a long vacation ;)
20:08 shyam joined #gluster-dev
20:08 raghu joined #gluster-dev
21:30 major so .. anyone have any objections to me rewriting all the ansible playbooks into 1 playbook that includes os/vendor vars to figure out package deps and such?
21:32 misc which playbooks ?
21:32 major tests/vagrant/
21:33 misc mhh
21:33 misc https://github.com/gluster/gluster.org_ansible_configuration/ the current build node config is here
21:33 misc in case you want to inspire and/or verify what we do
21:33 major basically was going to make it support centos6, centos7, ubuntu14, ubuntu16, fedora, debian, and everything else from a single playbook
21:33 misc https://github.com/gluster/gluster.org_ansible_configuration/tree/master/roles/jenkins_builder
21:34 misc major: I think that would be cleaner to avoid duplciation, so yes
21:35 major wow .. lotta playbooks in there
21:35 kkeithley @later tell nigelb I don't think the new build.sh got pushed to the netbsd regression boxes
21:35 glusterbot kkeithley: The operation succeeded.
21:36 major I am gonna focus on this one task for now :)
21:36 major in part because it is a minor annoyance in the tests I am trying to do
21:38 misc major: the other is the infra of gluster.org, in a ideal world, the vagrant one and the one in prod should be the same
21:38 major in an ideal world my editor would prevent me from writing copy/paste code :(
21:39 major still gotta start somewhere ..
21:39 major sides .. my ansible is rusty
21:42 misc if you have any questions, I will be happy to help
22:22 shyam joined #gluster-dev
23:55 raghu joined #gluster-dev

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