Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster-dev, 2016-07-11

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

All times shown according to UTC.

Time Nick Message
01:59 nishanth joined #gluster-dev
02:44 kdhananjay joined #gluster-dev
02:51 rafi joined #gluster-dev
03:20 magrawal joined #gluster-dev
03:39 nbalacha joined #gluster-dev
03:49 itisravi joined #gluster-dev
03:56 penguinRaider joined #gluster-dev
04:15 atinm joined #gluster-dev
04:23 atinmu joined #gluster-dev
04:25 MusiciAtin joined #gluster-dev
04:29 poornimag joined #gluster-dev
04:30 sanoj joined #gluster-dev
04:34 hgowtham joined #gluster-dev
04:39 ppai joined #gluster-dev
04:40 shubhendu joined #gluster-dev
04:46 ramky joined #gluster-dev
05:02 prasanth joined #gluster-dev
05:02 Manikandan joined #gluster-dev
05:06 ndarshan joined #gluster-dev
05:14 shubhendu joined #gluster-dev
05:15 sakshi joined #gluster-dev
05:20 Bhaskarakiran joined #gluster-dev
05:30 jiffin joined #gluster-dev
05:41 Muthu joined #gluster-dev
05:42 nishanth joined #gluster-dev
05:42 Saravanakmr joined #gluster-dev
05:50 pkalever joined #gluster-dev
05:52 hchiramm joined #gluster-dev
05:57 devyani7_ joined #gluster-dev
06:04 prasanth joined #gluster-dev
06:05 karthik___ joined #gluster-dev
06:07 Manikandan joined #gluster-dev
06:09 nigelb I've sent an email about this to gluster-infra@, but also announcing here: I've changed jenkins permissions for everyone.
06:09 nigelb You can only now trigger jobs and view them if you previously had access.
06:10 nigelb If there's something you do on a day-to-day basis that you cannot do anymore, please file a bug against project-infrastructure.
06:15 spalai joined #gluster-dev
06:17 kshlm joined #gluster-dev
06:19 aspandey joined #gluster-dev
06:21 asengupt joined #gluster-dev
06:22 aravindavk joined #gluster-dev
06:23 Bhaskarakiran joined #gluster-dev
06:26 nigelb The changes are documented here: https://github.com/gluster/infra-docs/pull/7
06:27 foster joined #gluster-dev
06:35 rafi1 joined #gluster-dev
06:36 rafi joined #gluster-dev
06:40 pranithk1 joined #gluster-dev
06:45 skoduri joined #gluster-dev
06:45 itisravi joined #gluster-dev
06:46 rafi1 joined #gluster-dev
06:51 ppai joined #gluster-dev
06:53 hchiramm_ joined #gluster-dev
06:58 msvbhat joined #gluster-dev
06:58 Manikandan joined #gluster-dev
07:03 nigelb misc: (for when you're awake) - what's on elk.cloud.gluster.org and who has access to it?
07:04 ndevos nigelb: that is ElasticSearch, LogStash, Kibana, I've been using it to get some statistics from Gerrit
07:04 ndevos but, it needs a lot of RAM, and always goes OOM...
07:05 nigelb ndevos: Hrm, I want to give it more RAM and get all our logs in one place.
07:05 nigelb How does it get data from Gerrit?
07:06 ndevos nigelb: I manually dump json files through logstash
07:06 misc nigelb: well, ndevos
07:06 nigelb Hrm.
07:06 misc (for the record, I am not a huge fan of having non official rpm stuff)
07:07 nigelb I may take over and make it a place to dump all our build logs.
07:07 nigelb I feel like centos CI gives us temporary machines.
07:07 nigelb Is that a correct guess?
07:07 ndevos the build logs are not so interesting
07:07 misc and we have a central syslog installation for system log, that should work
07:08 ndevos I fed it the logs from the regression tests at one point, but that really needs more resources for the system
07:08 ndevos we only capture those regression test logs for failed runs, and they are still ~200MB in size, I think
07:08 nigelb oh wow.
07:09 nigelb I didn't know the logs were that noisy.
07:10 ndevos unfortunately... but that is also because we test a lot of the failure/recovery paths, and restart the processes a lot
07:11 ndevos I would prefer to see some real world use-cases, maybe download.g.o, gerrit backups, jenkins artifacts, .. on gluster volumes
07:11 ndevos those logs should then have some more useful information, and we can check what messages get logged (too) much
07:12 nigelb ndevos: The question I have though is, if we move things to centos CI, will we still have logs on the machine like we currently do?
07:13 nigelb Or should we dump it in a centralized location?
07:13 ndevos nigelb: we can dump them somewhere, similar to how the nightly builds put the RPMs on artifacts.ci.centos.otg
07:13 ndevos *.org
07:14 nigelb aha
07:14 misc ndevos: mhh, what do you mean real wordl use case ?
07:14 nigelb I'm contemplating moving our rpm builds to centos CI to begin with.
07:14 nigelb and streamline the centos machines on rackspace.
07:14 misc (cause I did propose to have 2 redundant mirrors for download, but the consensus was that this was extra complexity)
07:15 hgowtham_ joined #gluster-dev
07:15 ndevos misc: regression tests are not really a use-case, and the logged results will be different from normal usage
07:16 ndevos misc: if the contents of the downloads, or other infrastructure bits are on Gluster, we could use the logs from that environment to see where we can improve
07:17 ndevos nigelb: we have scripts to build RPMs in the CentOS CI, it should not be difficult to modify those to also get triggered by Gerrit
07:17 misc ndevos: it may be the lack of coffee and/or lack of sleep, but that would be read only mostly, will that be that interesting ?
07:18 ndevos misc: yes, that is still interesting, it should be pretty quiet in that case, and we would want to know if that changes
07:18 nigelb ndevos: brilliant.
07:19 nigelb I want to convert them all to JJB
07:19 nigelb and update automatically with Jenkins.
07:19 misc ndevos: but so, we would run the stable gluster release, isn't it too late by then ?
07:19 ndevos nigelb: yeah, thats fine, I've just never used JJB, but it was suggested by others already too
07:20 nigelb It's way more readable :)
07:20 ndevos misc: maybe, but maybe we could run the latest release, update to the new 3.x every 3 months
07:21 nigelb https://github.com/gluster/glusterfs​-patch-acceptance-tests/blob/master/​jenkins/jobs/glusterfs-devrpms.yml
07:21 nigelb That's how it looks if you're curious
07:21 nigelb There's a lot of things templated out so it looks simpler.
07:22 misc ndevos: do we have a document with the lifecycle of each version and stuff like this, so I can try to see if that wouldn't cause some issue ?
07:23 misc (and the impact of a upgrade, etc, like do we need to stop download for that, or can it be done online, do they need to be in sync or not, etc)
07:24 nigelb I'm not entirely sure if we want to run gluster on production yet.
07:25 aravindavk joined #gluster-dev
07:25 nigelb what's the easiest way to virtualize on fedora? kvm? virtualbox?
07:25 nigelb I tried virtualbox. Doesn't seem to want to work easily.
07:25 misc kvm
07:25 misc you can take a look at vagrant
07:26 nigelb that's what we have on formicary, right?
07:26 misc or use stuff like virt-builder or virt-install
07:26 mchangir joined #gluster-dev
07:26 misc nigelb: yep
07:26 nigelb I'm trying to virtualize netbsd7.
07:26 nigelb It's going to be fun!
07:26 misc there is only 3 "big" choices on Linux, kvm, xen and virtualbox
07:26 misc virtualbox requires out of the tree kernel, so that's a pain on fedora
07:27 misc kvm is where RH put efforts, and xen is what (historically) rackspace used
07:27 post-factum vbox is pita on every system as those 3rd-party modules are awesome sh*t
07:28 nigelb I started with vbox and hit the out of tree kernel problem.
07:28 nigelb so kvm it is.
07:28 post-factum nigelb: just stick to kvm and be happy
07:28 misc post-factum: well, on windows, it seems to be ok :)
07:28 post-factum misc: no other option on windows, yeah. but who uses windows ;)
07:28 nigelb i can run this install command and go to lunch :)
07:28 nigelb I hadn't heard the fans spinning on this laptop until today.
07:29 nigelb clearly I wasn't using it well enough.
07:30 misc mhhh virt-builder --list do not show netbsd :(
07:33 rafi joined #gluster-dev
07:34 misc mhh https://jedipunkz.github.io/blog​/2013/03/28/netbsd-on-openstack/
07:36 aspandey joined #gluster-dev
07:40 aravindavk joined #gluster-dev
07:44 nigelb after a bit of tinkering, I think it's installing.
07:52 Manikandan joined #gluster-dev
07:53 hchiramm__ joined #gluster-dev
07:58 Bhaskarakiran joined #gluster-dev
08:00 skoduri_ joined #gluster-dev
08:02 misc nigelb: so for jenkins access control, what about using freeipa ?
08:02 misc (for having 1 single password, and/or 1 single ssh keys out there)
08:03 asengupt joined #gluster-dev
08:04 nigelb I believe we barely even need that. I think we may not even need freeipa.
08:04 Acinonyx joined #gluster-dev
08:04 misc well, if we do not want to log as root directly on server as you wanted, that's kinda needed :)
08:05 nigelb Not really. We can do that with just ansible alone.
08:05 misc true
08:05 misc so I guess I installed it for nothing...
08:06 nigelb Jenkins is going in the direction of "Don't really need access, we'll automate it against gerrit instead"
08:06 nigelb and Gerrit uses github already.
08:06 misc we have (future) monitoring
08:07 misc or rather, the plan is also to have monitoring
08:08 nigelb I'm wondering if we need ldap for that. Only a few people will need to ack alerts. That can also be done with ansible. The rest can be public or pseudo public with a shared common password.
08:08 misc that requires to distribute a password per user
08:09 nigelb ideally, I see only you or other osas sysadmins needing to ack alerts.
08:09 nigelb The rest of us just need to see the status.
08:09 nigelb we can all use a guest account which lets view status
08:09 nigelb Or do you see us all having accounts and ability to ack?
08:10 misc I was counting on having all people with root access to ack, yes
08:10 nigelb alternate idea for that is to have a nagios bot that can take acks over IRC via private channel.
08:11 misc chatops, seems doable, but before improving the monitoring, we need to have one :)
08:11 nigelb Indeed.
08:12 misc we also have wordpress that requires user account
08:14 itisravi_ joined #gluster-dev
08:24 asengupt joined #gluster-dev
08:24 nigelb I didn't realize you could use freeipa with wordpress.
08:25 misc that's ldap
08:25 misc and the whole point is to use it with all we can
08:26 misc I usually also try to delegate website building to people, but we can do that with jenkins too
08:33 itisravi_ joined #gluster-dev
08:34 aspandey joined #gluster-dev
08:38 nigelb I still think we can get further along in generally making our infra better before we bring out freeipa.
08:38 nigelb My solution to a lot of the access issue is "can we do something so you can do what you want without access".
08:39 pranithk1 joined #gluster-dev
08:44 atalur joined #gluster-dev
08:49 hgowtham__ joined #gluster-dev
08:49 Bhaskarakiran joined #gluster-dev
08:51 rastar joined #gluster-dev
08:54 hchiramm_ joined #gluster-dev
08:59 misc well, sure, but the fact is that freeipa is already there since 6 or 8 months :)
09:00 misc and when i said "we need to have a process before giving access", people said "yes", then didn't care
09:08 post-factum shouldn't http://review.gluster.org/#/c/14785/ be backported to 3.7 as well? we observe "failed to set TCP_USER_TIMEOUT -1000" warning sometimes
09:08 Manikandan joined #gluster-dev
09:09 ndevos post-factum: yeah, we can do that
09:10 post-factum ndevos: would be nice, thanks
09:26 nigelb ndevos: can you give me credentials to that ELK machine? Just to see the stats.
09:27 kotreshhr joined #gluster-dev
09:28 ndevos nigelb: I'm not sure if its still working... you want the webui?
09:29 nigelb yeah
09:29 nigelb just curious to see what stats you're collecting :)
09:31 ndevos well, I get a lot of "Courier Fetch: 116 of 1265 shards failed." and diagrams that tell me "no results found"
09:31 ndevos nigelb: https://devos.fedorapeople​.org/tmp/gerrit-on-elk.png
09:31 ndevos and https://devos.fedorapeople.org/t​mp/patchcounts-per-developer.png
09:32 msvbhat joined #gluster-dev
09:32 ndevos those are older images, but gives the idea
09:32 nigelb aah.
09:32 nigelb that is pretty good.
09:34 ndevos nigelb: I can create a kibana account for you, but it's pretty useless atm
09:34 nigelb ndevos: Not yet.
09:34 nigelb I might just rebuild that once we have the cage.
09:34 nigelb with a cron to collect data
09:36 nigelb ndevos: I've been trying to kill unused machines on rackspace. So I was curious what this was.
09:37 ndevos nigelb: we could kill it, but I'd like to copy some of the scripts that I have from it first :)
09:37 nigelb well, obviously. I'm not going to kill it right away.
09:37 nigelb I'll let you know before I kill it.
09:37 nigelb Or even plan to kill it.
09:38 ndevos ok :)
09:38 Manikandan joined #gluster-dev
09:39 Jules-2 joined #gluster-dev
09:41 misc nigelb: I kinda already cleaned old VM on rackspace
09:41 nigelb I'm also killing off machines reserved for Jenkins jobs that aren't run anymore.
09:42 nigelb so there's at least another 3 to 4 machines to kill.
09:42 nigelb And a round of netbsd resurrections :)
09:42 nigelb too many broken machines :(
09:42 * misc also need to see where the new server are
09:45 nbalacha joined #gluster-dev
09:46 misc nigelb: you have a ready made .travis.yml for the docs ?
09:46 nigelb misc: not for infra docs.
09:46 atinm joined #gluster-dev
09:46 nigelb feel free to make it :)
09:46 misc well, yeah, I want, but if I can just cut and paste :)
09:46 nigelb you probably just need a make html
09:46 misc nigelb: because while I see why you would do a PR and merge it on your own, it seems a bit strange
09:47 nigelb Oh, I just merged off all the pending PRs today.
09:47 nigelb they were mostly spelling fixes ^_^
09:50 nishanth joined #gluster-dev
09:53 kshlm ndevos, Those kibana gerrit stats are cool! They look better than bitergia.
09:53 karthik___ joined #gluster-dev
09:53 kshlm Where have you been hiding them till now?
09:54 ndevos kshlm: its just that the system is not stable enough to use :-/
09:54 hchiramm__ joined #gluster-dev
09:55 kshlm nigelb, The devrpm links being reported back to gerrit are 404ing.
09:56 nigelb oh dear.
09:56 nigelb kshlm: link?
09:56 kshlm For eg., https://review.gluster.org/14224
09:56 kshlm The actual links https://build.gluster.org/job/devrpm-fedora/149/
09:56 kshlm https://build.gluster.org/job/devrpm-el7/144/
09:56 kshlm https://build.gluster.org/job/devrpm-el6/142/
09:56 nigelb I bet that's my access stuff
09:58 kshlm The build history for those jobs stop at about 10 entries.
09:58 ndevos post-factum: here you go: http://review.gluster.org/#/q/Ibc162​64ceac0e69ab4a217ffa27c549b9fa21df9
09:58 nigelb OH.
09:58 nigelb that's intentional.
09:58 nigelb I recently changed that from 50 to 10.
09:58 kshlm Why?
09:59 post-factum ndevos: cool and thx
09:59 nigelb I was talking to kkeithley (I think) or ndevos and we didn't think it was actually being used.
09:59 nigelb Let me fix that up right away.
09:59 karthik_ joined #gluster-dev
09:59 kshlm I'd like to know what failed.
09:59 ndevos nigelb: the RPMs are not used for all I know, but the buildlogs would be
09:59 nigelb Right, now I realize that.
09:59 kshlm Could jenkins be set up to archive failed build logs?
10:00 nigelb I can't do anything about old jobs.
10:00 nigelb But I can make sure we go back to 50 days.
10:00 nigelb pushing up the change now.
10:09 misc kshlm: we can surely, but we need to plan for space
10:10 misc (I assume we do not need backup of the failed build ?)
10:17 post-factum ndevos++
10:17 glusterbot post-factum: ndevos's karma is now 284
10:17 post-factum ndevos: one would start trading his karma
10:18 post-factum ndevos: e.g., 10 points for 1 vacation day
10:23 nigelb We really should put a https://download.gluster.org/pub/gluster/​glusterfs/LATEST/RHEL/glusterfs-epel.repo to centos SIG repo :)
10:23 nigelb I bet it's part of many folks' scripts.
10:24 kdhananjay1 joined #gluster-dev
10:24 aspandey joined #gluster-dev
10:27 hchiramm__ joined #gluster-dev
10:30 Acinonyx joined #gluster-dev
10:32 nigelb FYI - read the docs is down. In case someone asks.
10:32 aspandey joined #gluster-dev
10:33 kdhananjay joined #gluster-dev
10:36 Manikandan joined #gluster-dev
10:46 Saravanakmr joined #gluster-dev
10:51 nigelb misc: what's on gluster-sonar?
10:52 Acinonyx joined #gluster-dev
10:54 nishanth joined #gluster-dev
10:58 itisravi joined #gluster-dev
10:59 Manikandan joined #gluster-dev
11:06 atinm joined #gluster-dev
11:11 ashiq joined #gluster-dev
11:20 rastar joined #gluster-dev
11:23 Manikandan joined #gluster-dev
11:25 Manikandan joined #gluster-dev
11:27 hchiramm_ joined #gluster-dev
11:42 rraja joined #gluster-dev
11:43 hgowtham_ joined #gluster-dev
11:44 pkalever left #gluster-dev
11:47 kotreshhr left #gluster-dev
11:48 ppai joined #gluster-dev
11:48 Muthu joined #gluster-dev
11:49 kotreshhr joined #gluster-dev
11:49 Manikandan joined #gluster-dev
11:53 ira joined #gluster-dev
11:57 anoopcs ndevos, Have a you already sent a patch for http://www.gluster.org/pipermail/g​luster-devel/2016-July/050035.html ?
11:58 ndevos anoopcs: building a test-package now to see if my patch works
11:58 anoopcs ndevos, Why do you need to build a test package to test it out?
11:59 ndevos anoopcs: to verify if it actually works as intended, or if I missed something
12:00 ndevos anoopcs: actually, I almost always test the changes I make... mostly by building RPMs and running them
12:00 ndevos anoopcs: http://termbin.com/k5es is the patch I came up with
12:01 ndevos and that seems to do it too ):
12:01 ndevos uh, :)
12:04 Saravanakmr ndevos, nigelb  There is a weird issue - one patch is reporting smoke test error due to another bug 1349657
12:04 glusterbot Bug https://bugzilla.redhat.com:​443/show_bug.cgi?id=1349657 low, unspecified, ---, johnzzpcrystal, MODIFIED , process glusterd set TCP_USER_TIMEOUT failed
12:05 Saravanakmr ndevos, I can see you have changed the status of 1349657 to MODIFIED :)
12:05 Saravanakmr ndevos, nigelb  http://review.gluster.org/#/c/14830/   - https://build.gluster.org/job/compare-b​ug-version-and-git-branch/19263/console
12:06 ndevos Saravanakmr: yeah, the patch for that has been merged?
12:06 Saravanakmr ndarshan, yes it is merged
12:06 Saravanakmr ndevos, yes it is merged but seeing issue in another patch
12:07 Saravanakmr ndarshan, sorry..ignore
12:07 ndevos Saravanakmr: hmm, right, the other patch is for bug 1351071
12:07 glusterbot Bug https://bugzilla.redhat.com:​443/show_bug.cgi?id=1351071 medium, unspecified, ---, sarumuga, ASSIGNED , [geo-rep] Stopped geo-rep session gets started automatically once all the master nodes are upgraded
12:07 Saravanakmr ndevos, weird !
12:08 ndevos Saravanakmr: I have no idea why that could happen :-/
12:08 Saravanakmr nigelb, can you please have a look?
12:09 Saravanakmr nigelb, ndevos another patch from Aravinda also fails with same reason http://review.gluster.org/#/c/14627/
12:11 ndevos Saravanakmr: oh, that actually is the same commit-id, and the commit message is used to parse the BUG:
12:12 misc nigelb: a sonarqube instance managed by Louis Zuckerman
12:12 skoduri One of my patches too ran into same issue  - https://build.gluster.org/job/compare-b​ug-version-and-git-branch/19265/console
12:12 skoduri nigelb, ndevos , Saravanakmr ^^^
12:13 ndevos skoduri: is that also doing a "Checking out Revision b2c73cbf423de6201f956f522b7429615c88869d (origin/master)" ?
12:13 mchangir joined #gluster-dev
12:13 skoduri ndevos, yes
12:14 skoduri ndevos, and from http://review.gluster.org/#/c/14878/3, looks like that was parent commiID
12:14 ndevos hmm, maybe nigelb updated the Jenkins job to the new Jenkins Job Builder configuration and some option or something was missed?
12:15 skoduri okay
12:17 ndevos obviously there was an update to the new .yml file 2 hours ago, not sure if that goes live automatically
12:18 ndevos https://github.com/gluster/glusterfs-patc​h-acceptance-tests/blob/master/jenkins/jo​bs/compare-bug-version-and-git-branch.yml for the curious, but I have no idea what options are needed or how it releates to other config
12:20 ndevos nigelb: the compare-bug-version-and-git-branch script expects the to-test change to be checked out, not sure if the JJB config does that?
12:28 hchiramm__ joined #gluster-dev
12:45 anoopcs v3.7.13 is missing in bugzilla.
12:58 lpabon joined #gluster-dev
13:01 kkeithley file against 3.7.12 for now.  3.7.13 will be there soon (ditto for 3.8.1)
13:08 ira joined #gluster-dev
13:18 atinm joined #gluster-dev
13:21 atinmu joined #gluster-dev
13:22 spalai joined #gluster-dev
13:28 hchiramm_ joined #gluster-dev
13:28 ndevos kkeithley: I really wonder if we need such detailed versions in bugzilla, we ask users for the exact version in the comment already too
13:29 ndevos I would be happy with just 3.6, 3.7, 3.8 and mainline
13:29 ndevos maybe even rename mainline to devel
13:31 nigelb ndevos / Saravanakmr - looking
13:31 nigelb sorry I was commuting back.
13:31 kkeithley I'd be okay with it.  Don't you think there should be some little bit of discussion, either on the mailing list or in the Community Meeting?
13:31 pranithk1 joined #gluster-dev
13:31 kkeithley How do we enforce version in the comment?
13:35 kkeithley s/enforce/ask for/
13:44 ndevos kkeithley: yes, we should ask about it in the community meeting, or maybe just on the maintainers list
13:44 ndevos kkeithley: we triage bugs, and if the version is important enough to know, we should ask for it in case it is missing - but most add it already
13:45 nigelb Saravanakmr: It should be fixed now
13:45 Saravanakmr nigelb, Thanks!
13:45 Saravanakmr nigelb++
13:45 glusterbot Saravanakmr: nigelb's karma is now 13
13:45 nigelb I've retried all your jobs
13:45 nigelb Gerrit should report success
13:46 Saravanakmr nigelb, ok
13:48 nigelb bug 1354529 explains what happened if you're curious
13:48 glusterbot Bug https://bugzilla.redhat.com:​443/show_bug.cgi?id=1354529 unspecified, urgent, ---, nigelb, ASSIGNED , compare-bug-version-and-git-branch failing for seemingly no reason
13:49 shyam joined #gluster-dev
13:53 nbalacha joined #gluster-dev
14:12 ndevos nigelb: time for a quick review of https://github.com/gluster/gluster​fs-patch-acceptance-tests/pull/35 before I activate that?
14:19 ndevos nigelb: I guess not, and I'll merge it now :)
14:20 * nigelb looks
14:20 nigelb This XML hurts to read.
14:21 nigelb ndevos: I'll change all of these over to yml soon :D
14:21 nigelb Once I figure out how I messed up compare-bug
14:21 ndevos nigelb: I understand the XML better than the yml :)
14:22 ndevos nigelb: compare-bug expects the git repository to have the HEAD commit the one to test, I think it checks out the master branch/HEAD instead
14:22 ndevos nigelb: or, didnt you mention it got fixed?
14:25 jiffin joined #gluster-dev
14:27 nigelb ndevos: I reverted to the old commit.
14:27 kkeithley Well, that's just weird.  We've been building 3.7.x pkgs on Debian wheezy just fine, using pkgconfig to check for liburcu-bp, even though liburcu-dev doesn't provide pkgconfig files.
14:27 ndevos huh, magick!
14:27 kkeithley But now 3.7.13 doesn't want to build any more.
14:29 nigelb ndevos: I haven't actually fixed the issue in the JJB yaml yet.
14:29 ndevos and I was building epel-7 packages just fine until a few minutes ago, and now I hit the 'missing ....pyc/pyo files' that was reported before
14:29 hchiramm__ joined #gluster-dev
14:29 ndevos nigelb: ah, ok, yes, reverting that part would do it too :)
14:30 kkeithley hmmm.   I put those patches on the back burner because it stopped happening.
14:31 nigelb kkeithley: would you be averse to me version controlling package building at some point?
14:32 nigelb (I don't have bandwidth to do this quite yet, but I do want to make your life easier.)
14:32 ndevos nigelb: arent they version controlled in pkgs.fedoraproject.org and similar packaging infrastructures?
14:33 kkeithley package building already has version control.   The various RPM .spec files and Debian packaging bits are all in git
14:33 nigelb Oh. I hadn't realized that.
14:33 kkeithley Or is there something else you're proposing?
14:33 nigelb My end goal is for you to be able to run a jenkins job to build all packages from commit X
14:34 nigelb long long way away.
14:34 ndevos how can we tie that in with Fedora Koji+Bodhi, the CBS, Ubuntu, ... etc?
14:36 kkeithley sounds like a nice idea.  Every time I build pkgs I find some kind of breakage that requires non-trivial work to fix up.
14:37 kkeithley Might be a $hard_problem
14:39 kkeithley might be nice to fire off period package builds in, e.g., Ubuntu Launchpad and SuSE Build System, just to find the breakage before a release
14:40 nigelb that's exactly what I'm thinking.
14:40 ndevos there is work being done in the CentOS CI to allow building packages in the CBS :D
14:40 aravindavk joined #gluster-dev
14:44 kkeithley s/period/periodic/
14:46 kkeithley Anyway, Debian & Ubuntu packaging bits are in https://github.com/gluster/glusterfs-debian
14:46 lpabon_ joined #gluster-dev
14:47 kkeithley SuSE rpm bits are in https://github.com/gluster/glusterfs-suse
14:48 kkeithley Fedora rpm bits are in Fedora dist git at pkgs.fedoraproject.org/glusterfs
14:49 nigelb aha
14:49 nigelb I'll tuck that into my notes.
14:49 kkeithley Ubuntu builds are done in Launchpad, which requires a shared secret that's kept on an internal server
14:50 nigelb Oh yeah, I know launchpad quite well.
14:50 nigelb Even contributed code it back in the day.
14:50 kkeithley SuSE builds are done in (my) SuSE build service account.   We could think about having a community account instead.
14:51 kkeithley Fedora and "legacy" EPEL builds are done in Koji.  Starting with 3.8 CentOS/EPEL builds are done in CBS.
14:54 kkeithley We do 3.7 (and 3.6 IIRC) too in CBS.  Point being that we're moving EPEL from Koji to CBS.
14:54 nigelb Noted.
14:54 nigelb The Ubuntu bits are easiest to automate, I believe.
14:54 nigelb I remember they did some work to do precisely this.
14:54 kkeithley Launchpad is very scriptable.
14:56 nigelb Just some fun in doing git -> bzr conversion.
14:56 rafi joined #gluster-dev
15:00 nigelb misc: where are the results of this sonarqube run?
15:00 nigelb er, semiosis ^
15:00 misc nigelb: no idea :/
15:01 nigelb misc: that wasn't meant for you, sorry!
15:02 pkalever joined #gluster-dev
15:06 wushudoin joined #gluster-dev
15:11 shyam joined #gluster-dev
15:11 kkeithley oh?  I just tried redoing the 3.7.12 build on wheezy (pbuilder --build glusterfs_3.7.12.dsc) and that still works, finds liburcu-bp just fine.
15:12 kkeithley git blame says the pkgconfig test in configure.ac hasn't changed since kshlm added it back in 2015
15:13 nigelb This is strange.
15:14 nigelb I created a new job to test compare-bug
15:14 nigelb and it works flawlessly.
15:14 kkeithley well, we can't blame solar flare activity, because there isn't any.
15:14 owlbot joined #gluster-dev
15:15 nigelb I'm going to blame looking at it late at night. One for the morning.
15:38 ashiq joined #gluster-dev
16:05 rafi joined #gluster-dev
16:08 shubhendu joined #gluster-dev
16:12 devyani7_ joined #gluster-dev
16:19 semiosis nigelb: hi!
16:21 semiosis nigelb: the only sonarqube i'm familiar with is the one i set up for glusterfs-java-filesystem.  is that what you're talking about?
16:23 nigelb semiosis: yep!
16:23 semiosis nigelb: how can I help you?
16:23 nigelb semiosis: do we use the instance?
16:23 nigelb (I'm cleaning up rackspace machines pretty thoroughly)
16:25 semiosis you can decommission it
16:25 nigelb excellent, thank you.
16:25 semiosis it hosted code quality reports for this project, https://github.com/semiosis​/glusterfs-java-filesystem, but the links are dead now.
16:26 semiosis thank you
16:30 jiffin joined #gluster-dev
16:47 kkeithley nigelb: rewinding back to source controlling packaging bits...   Maybe once things settle down you'd be interested in taking over a chunk of the packaging job. Debian, Ubuntu, SuSE, at least.  Maybe even Fedora too.  Is that something you'd be interested in?
16:50 devyani7 joined #gluster-dev
17:18 pranithk1 joined #gluster-dev
17:23 ira joined #gluster-dev
17:32 skoduri joined #gluster-dev
19:10 ira joined #gluster-dev
23:15 ira joined #gluster-dev

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