IRC logs for #baserock for Tuesday, 2014-10-07

*** genii [~quassel@ubuntu/member/genii] has quit [Read error: Connection reset by peer]00:44
*** radiofree [radiofree@unaffiliated/radiofree] has quit [Quit: Bye]03:24
*** radiofree [radiofree@unaffiliated/radiofree] has joined #baserock03:30
radiofreepedroalvarez: can you make that change in the morning?03:46
pedroalvarezYes. Anybody wants to review it? 07:22
*** mSher [] has joined #baserock07:31
*** dutch_ [] has joined #baserock07:46
pedroalvarezthis is the full diff07:51
pedroalvarezi have never updated a tarball lorry, so I don't know the consequences07:52
pedroalvarezreviews appreciated07:52
*** mSher [] has quit [Ping timeout: 250 seconds]07:56
jjardonIm getting "500 - Internal Server Error" in g.b.o, known issue?08:04
pedroalvarezI have just restarted the lighttpd-git service08:06
pedroalvareztry now08:06
pedroalvarezjjardon: ^^08:06
jjardonpedroalvarez: works! thanks ;)08:06
*** mSher [] has joined #baserock08:10
pedroalvarezfor future reference, what I've done:
*** tiagogomes [~tiagogome@] has joined #baserock08:21
* Kinnison boggles why it'd be saying it was out of memory08:36
*** jonathanmaw [] has joined #baserock08:47
*** ssam2 [] has joined #baserock09:15
jmacsSince I've made a new machine with 14.40, the instructions on don't seem to work any more09:29
jmacsAt deployment I get: ERROR: Couldn't find morphology: base-system-x86_64-generic.morph09:29
ssam2oho, those instructions are out of date09:30
ssam2thanks for pointing it out09:30
ssam2I'll update them09:31
*** franred [] has joined #baserock09:33
jmacsIs there anything quick I can do to get my deploy working in the meantime?09:34
straycatprepend systems/ to the morph: line09:34
ssam2jmacs: this is the fix, I've done it now:;a=blobdiff;f=devel-with.mdwn;h=acaf74b6d9ad9014d1cd375a912449d2fa0a9472;hp=3417793b1dd37c81cec935028541c4296fcbb87a;hb=ed78a7afd7726b04d5da4864ed035497020d0fec;hpb=875eeac8fdeeeb84f73e0ac8e70e7323451868f409:34
wikicatssam2: Wiki change: Fix example cluster morphs to work with Baserock 14.40 release
wikicatssam2: Wiki change:
ssam2right, thanks09:35
wikicatWiki change: Fix example cluster morphs to work with Baserock 14.40 release
jmacsYep, that works09:35
straycatwhat? why did it reply to ssam2 directly?09:36
KinnisonPerhaps it saw the SHA1s?09:38
* straycat nods09:38
wikicatstraycat: Wiki change: Fix example cluster morphs to work with Baserock 14.40 release
* straycat didn't know about this09:39
radiofreepedroalvarez: i can give it a +2 if it'll get it merged :)09:55
straycatDoesn't that replace the old version of boost? What if someone is using that?09:56
radiofreeis there any way i can expend my /src/ direction in a vm?10:00
radiofreei used the fallocate method10:00
richard_mawradiofree: if by expending your /src/ direction you meant expand your /src/ disk, then yes. I use `truncate -s +2G $DISK_IMAGE`, shut down the VM, start it up again and do a `btrfs filesystem resize max /src`10:01
pedroalvarezradiofree: done
radiofreeawesome pedroalvarez!!10:03
* radiofree might have enough time to build that10:03
radiofreerichard_maw: thanks10:03
radiofreeactually i'm out of space in /home now10:03
radiofree(on the host)10:03
* richard_maw scrolls back further10:03
richard_mawjjardon: woo on building systemd master10:04
pedroalvarezstraycat: lorry replaces master, but it keeps tags with every tarball version and also it keeps the branches10:04
pedroalvarezI've checked what happened with the gcc-tarball before doing the push10:05
radiofreerichard_maw: do i run truncate -s +2G $DISK_IMAGE when the vm is running?10:07
richard_mawjjardon: the network stuff is to a. remove our systemd units and udev rules for running ifup from the busybox chunk, and b. work out how networkd works10:08
richard_mawradiofree: you could do, but I don't think your VM will notice the disk has increased in size until you reboot10:08
richard_mawthough I don't know if this is the case if we've got hotplug10:08
ssam2richard_maw: surely touching a disk image that's in use in a VM from outside the VM will result in corruption ?10:12
ssam2truncate may need to completely move the data, surely ?10:12
ssam2radiofree: to be safe, I'd power off your vm *then* run `truncate`10:13
straycatpedroalvarez, Oh okay that's fine then.10:15
richard_mawssam2: I'm fairly sure kvm's file-backed disk images work high enough in the stack that it can handle the data being physically moved in your disk10:17
richard_mawtruncate in extend mode will only say that the file is now larger10:17
ssam2richard_maw: that requires them to have implemented an inotify handler to detect that the disk moved. Which they may have done, but not necessarily10:26
ssam2I guess with truncate it's probably OK if it doesn't change anything on disk, I don't really understand how that would work though10:27
richard_mawssam2: I don't follow why inotify is needed, it works with an open fd10:27
ssam2oh, right.10:28
richard_mawit might need inotify to notice whether the file's gotten larger, but increasing its size shouldn't break anything10:28
ssam2ok. having broken vm disk images in the past by having them mounted in two different places, I try to be careful about this stuff10:29
richard_mawhaving them mounted in two different places is a different problem, caused by 1. filesystems assuming they have exclusive access to the disk and 2. there being no suitable synchronisation primitive that wouldn't be slow, even if a fs decided to allow multiple systems to concurrently mount it10:31
richard_mawpedroalvarez: your cloud-init.configure patch has +3 now, do you want to merge it, or should I?10:32
pedroalvarezrichard_maw: I will do it, thanks :)10:33
*** fay_ [] has joined #baserock11:03
jjardonseems https access to g.b.o doesnt work: "fatal: unable to access <repo>: SSL certificate problem: self signed certificate11:32
jjardonrichard_maw: ok, I will take a look11:33
jjardonrichard_maw: BTW, Do I still have your +1 for the "move llvm" and "move libdrm" branches after the fixes I made? (basically build llvm 3.3.1 instead 3.5)11:35
pedroalvarezgerrit instance with the baserock projects:
CTtpollardwhere is this instance running pedroalvarez?11:58
pedroalvarezin a public accesible cloud11:59
pedroalvarezit is *not* running on baserock11:59
CTtpollardcool :)12:00
richard_mawjjardon: you need to set GIT_SSL_NO_VERIFY=1 when cloning https unless you've got a really new version of baserock12:08
richard_mawjjardon: If the changes are as trivial as changing versions, yes, but I'm afraid I don't really follow what's going on in the graphics stack12:09
jjardonrichard_maw: Im trying to clone from my laptop12:10
pedroalvarezthe problem is: self signed certificate12:11
pedroalvarezwow, anybody can send patches to the gerrit instance12:14
KinnisonDepends on how gerrit is configured12:15
pedroalvareztrue, but I mean, it's possible!12:16
pedroalvarezI managed to login using yahoo, and also launchpad openid12:17
Kinnisonmob pushing is quite possible, yes12:17
pedroalvarezyou are all welcome to play with it :)12:17
franredit was possible before, but I think you should announce it as a gerrit test12:18
richard_mawI see it's only got our baserock/ project repositories, rather than lorries and delta etc.12:18
pedroalvarezrichard_maw: yeah12:19
pedroalvarezfor information about how to use gerrit, I've followed this instructions:
pedroalvarezI guess there are better tutorials out there12:24
*** fay [] has joined #baserock12:29
fay is now known as Guest7474212:29
*** fay_ [] has quit [Ping timeout: 260 seconds]12:32
*** thecorconian [~thecorcon@] has joined #baserock12:35
jjardonpedroalvarez: nice! But what are the advantages of using gerrit instead the mailing list? Does gerrit do some integration work? ie, build a image when you send a patch to verify nothing breaks?12:35
Guest74742 is now known as fay_12:36
Kinnisonit does proper patch tracking and can act as an aggregator of votes, track multiple versions of a patch, etc.12:36
pedroalvarezjjardon: you can send patches for review easily, and the team can vote the patch and merge things12:36
KinnisonMerging will not be handled by gerrit12:36
Kinnisonit'll be handled by something which talks to gerrit12:36
CTtpollardWe'll be able to set it to only accept an in-house OpenID provider12:43
CTtpollardfor anyone interested I'm currently trying to switch-out the default 'h2' db to postgres, I've succesfully done it manually on baserock, not as much luck doing it automated during morph build 12:45
pedroalvarezCTtpollard: it doesn't need to be done when building. 12:48
pedroalvareze.g., when we deploy a new Trove, all the configuration happens on first boot12:48
CTtpollardI've been looking into how it was done with gitlab, which seems to have postgres configed after the fact through the gitlab-setup.configure file 12:50
CTtpollardit should be able to do within the gerrit chunk morphology however I think12:50
KinnisonIf the act of doing so creates the postgres db etc, then you can't do it until the system boots12:51
franredKinnison, why not?12:51
Kinnisonbecause until deploy time, users might not exist, and you can't run system-code at deploy time12:52
Kinnisonso you have to do database setup after boot12:52
franredKinnison, you can create the users, run postmaster, add the configuration on integration time and then create a service which runs your postgres later12:53
franredwe have done this with gerrit12:54
Kinnisonhmm, I suppose system-integration time is okay12:54
ssam2pedroalvarez: -- log of deployment failure due to cloud-init.configure, from a Mason instance that's running at Codethink13:38
pedroalvarezthe attempt of upgrading mason with distbuild fixes is failing: The controller log says:
ssam2that line should be 'event.artifact.source.cache_key'14:27
ssam2as should anywhere else that 'artifact.cache_key' is accessed14:28
ssam2if that was missed, there may be other places that were missed too14:28
pedroalvarezI hacked that in the instance, waiting for results14:32
richard_mawbah! I thought I caught them all14:34
straycatrichard_maw, I didn't realise you'14:35
straycatd needed to modify the worker build scheduler?14:35
pedroalvarezthere were another one in that file, distbuild seems to work now14:44
richard_mawthose are the only two remaining in distbuild, the only remaining uses after that are in the list-artifacts command, which doesn't have any tests, so I didn't notice14:55
straycatWhy are you needing to touch the worker build scheduler?15:01
richard_mawbecause it does stuff with the cache id15:03
richard_mawuses it for build IDs I think15:03
straycatThat doesn't really help me understand why you need to make changes there, maybe when you resubmit your distbuild fixes it will become clear.15:04
richard_mawI'm not sure what more there is to say. It used to use artifact.cache_key for things, but now the cache_key is in the Source, so you need to use artifact.source.cache_key.15:08
straycatAhh well I didn't know the cache key had moved :)15:09
*** genii [~quassel@ubuntu/member/genii] has joined #baserock15:16
* richard_maw wants VLAN_8021Q in our kernels16:07
richard_mawI may provide patches later16:07
KinnisonShould be simple enough to add16:07
*** dutch_ [] has quit [Quit: Quit]16:09
*** jjardon [sid723@gateway/web/] has quit [Remote host closed the connection]16:21
*** jjardon [sid723@gateway/web/] has joined #baserock16:41
*** mSher [] has quit [Quit: Leaving]16:55
*** jonathanmaw [] has quit [Quit: Leaving]17:00
*** tiagogomes [~tiagogome@] has quit [Quit: Leaving]17:02
* jjardon running weston in systemd master in his vm. Probably network is broken but at least everything else seems to work17:05
*** ssam2 [] has quit [Quit: Leaving]17:17
franredwhat is the baserock cleanest way to add one repo into another? I have a repo A which is just a folder with some python scripts but nothing to install them, and other repo needs it to compile. Add the repo A to some folder and add it to the path when I will compile (call make) the repo B will work?18:16
radiofreegbo is down?19:46
radiofree500 - Internal Server Error19:46
radiofreeit's back :D19:48
pedroalvarezHas anybody fixed it?22:11
pedroalvarezI shall look 22:12
pedroalvarezNope, it came back to life itself 22:15
*** vmeson [~quassel@] has quit [Ping timeout: 245 seconds]22:18
*** genii [~quassel@ubuntu/member/genii] has quit [Read error: Connection reset by peer]22:41
pedroalvarezha! cgit working with gerrit23:26
pedroalvarezand to access gerrit
pedroalvarezgood night! :)23:27
paulsherwoodw00t! :)23:32

Generated by 2.14.0 by Marius Gedminas - find it at!