/srv/irclogs.ubuntu.com/2015/01/05/#juju-dev.txt

thumperdavecheney: you around?00:25
thumperdavecheney: I'd like you to join waigani and me in a hangout00:25
thumperto debug this channel problem00:25
davecheneythumper: on the phone00:30
davecheneydo you have a working code sample ?00:30
thumperdavecheney: we have a non-working code sample00:31
davecheneythumper: still OTP00:33
davecheneycan you show me the code ?00:33
davecheneythumper: ok, off da phone00:38
davecheneywaigani: var timeout chan something // can't remember01:10
davecheneyif req.timeout != [01:10
davecheney timeout = time.After(timeout)01:10
davecheney}01:10
davecheneyselect {01:10
davecheneycase <- timeout: // will only be selectable if timeout != nil01:11
davecheneywaigani: thumper http://play.golang.org/p/PKlJ1dMvvM01:15
axwanastasiamac: hiya. do you want to have a hangout? not much to talk about yet... :)01:17
anastasiamacaxw: happy new year!01:17
anastasiamacaxw: m actually still on holiday today ;(01:18
axwanastasiamac: ah, no worries. get out of here ;)01:18
anastasiamacaxw: u r holding our tanzanite fort!01:18
axwhappy new year to you too01:18
anastasiamacaxw: thnx :) talk tomorrow!01:18
perrito666nites, happy new year to all01:19
perrito666anybody seen walyworld?01:19
anastasiamacperrito666: u2!!! :)01:19
anastasiamacperrito666: someone wished me a Happy New 2005! I was so happy for a while :)01:19
anastasiamacperrito666: wallyworld caming today (still on holidays) - back tomorrow01:20
anastasiamacperrito666: camping even01:20
perrito666ah ok, then Ill see him tomorrow afternoon for me, I am actually on holiday too, its sunday night here01:20
anastasiamacperrito666: have a gr8 sunday before work :)01:21
waiganidavecheney: thanks for the help and notes. I've jotted them down for the new branch.01:21
perrito666anastasiamac: tx, although there are only 1.5 hs left to my sunday01:21
thumperanastasiamac: wallyworld camps all the time ;-)01:24
thumperanastasiamac: and if you are on holiday, what are you doing here?01:24
perrito666ok fun ppl, see you all tomorrow01:29
perrito666tadaa01:29
=== negronjl_afk is now known as negronjl
davecheneywaigani: np02:09
jamhappy new year everyone!03:49
wwitzel3jam: happy new year :)03:50
jamwwitzel3: and to you, though why are you sitting in Juju-dev at Sunday Midnight ?03:53
wwitzel3jam: just lurking04:04
thumpero/04:36
thumperhappy new year to you both04:36
* thumper is done for the day04:36
jamsame to you thumper, have a good evening04:37
thumperciao04:37
dimiternmorning folks!06:19
jammorning dimitern06:24
dimiternhey jam06:29
TheMuemorning08:12
dimiternTheMue, morning08:52
TheMuedimitern: heya, and a happy new year.08:54
dimiternTheMue, thanks! and cheers to you too ;)08:54
voidspacemorning all, happy new year09:10
voidspacefwereade: o/ morning, happy new year09:24
=== ev_ is now known as ev
=== liam_ is now known as Guest28200
dimiternvoidspace, fwereade, hey, morning and happy NY as well ;)10:00
voidspacedimitern: morning10:00
fwereadedimitern, voidspace: happy new years!10:01
perrito666good morning everyone and happy new year to those I didn't see since last year11:12
=== coreycb` is now known as coreycb
voidspaceooh, maas update worked flawlessly13:05
voidspaceperrito666: morning, happy new year13:05
perrito666voidspace: dont sound so surprised :p13:06
voidspaceperrito666: I know about computers13:06
voidspaceperrito666: I'm constantly surprised that anything works at all13:06
katco happy new year all!14:19
perrito666idem katco14:20
katcoperrito666: i'm not familiar with that acronym :p14:20
perrito666lol, sorry that is latin I think, its the word used in spanish for likewise or ditto14:21
katcoperrito666: ah ok :)14:21
katcoperrito666: perhaps as in "idempotent"14:22
dimiternkatco, perrito666, hey ;) happy NY and welcome back14:22
perrito666idem means equals iirc, and in spanish, to say likewise we use a word which roughly translates to equally14:22
katcodimitern: same to you! welcome back!14:23
perrito666dimitern: hey happy new york to you too :p14:23
dimiternthanks!14:23
dimiternkatco, btw if you haven't tried goflymake in emacs (with flycheck or flymake) it's *awesome* - syntax checking on the fly running go build in the background14:23
katcodimitern: yes, i run it :)14:24
dimiternkatco, :) nice! - I wasn't sure it brings a lot, but now I really like it14:24
katcodimitern: yeah, i like it too14:24
=== cmars` is now known as cmars
=== ChanServ changed the topic of #juju-dev to: https://juju.ubuntu.com | On-call reviewer: see calendar | Open critical bugs: 1407699
perrito666wwitzel3: ericsnow getting there, for some reason my connection to plus.google.com is not working15:01
voidspacedimitern: so, it's worse15:07
voidspacedimitern: we need to list the nodegroups and then list the interfaces for all groups15:07
dimiternvoidspace, oh what's that?15:07
voidspacedimitern: it's only *marginally* worse15:07
dimiternvoidspace, yeah :/15:07
voidspacedimitern: it's n calls instead of one15:07
voidspacedimitern: and the interfaces call doesn't have the cidr15:08
voidspacedimitern: I can *construct* a CIDR I think15:08
dimiternvoidspace, worthy of a maas bug15:08
voidspacedimitern: but the output isn't immediately compatible with SubnetInfo15:08
voidspacedimitern: I have a bug, I can add to it15:08
dimiternvoidspace, it has static ip range high and low, right?15:08
voidspacedimitern: yep15:09
voidspacedimitern: http://pastebin.ubuntu.com/9677036/15:10
dimiternvoidspace, why do you need a cidr?15:11
dimiternvoidspace, ah, to match yeah15:11
dimiternvoidspace, well it should be trivial to construct one from "ip" + "subnet_mask"15:12
voidspacedimitern: nodegroups-interface doesn't have vlan tag either15:13
voidspacedimitern: and the provider id will be the nodegroups-interface id not the network id (I assume)15:13
voidspacedimitern: unless a nodegroup interface and a network are the same thing15:14
voidspacewhich is unclear to me15:14
voidspacedimitern: they're slightly different (and the names are different) based on the maas cli here15:15
voidspacethe nodegroup-interface has the name "eth0" whereas the network has the name "maas-eth0"15:15
dimiternvoidspace, nope - afaik interfaces and networks are different entities in the maas db15:15
voidspacewhich worries me slightly - does it *matter* which we use15:16
voidspaceI would expect it to matter15:16
dimiternvoidspace, so to get enough to fill in SubnetInfo you'll need to use both the network config and the interfaces15:16
dimiternvoidspace, for juju, the subnet's provider id is the maas network name15:17
voidspacedimitern: right but they have different names *etc*15:17
voidspaceso matching them up is hard15:17
voidspacein my local maas the network has ip 172.16.0.0, whereas the nodegroup interface has the ip 172.16.0.215:17
dimiternvoidspace, istm you can only match them if the network's cidr and the interface's generated cidr are the same15:17
dimiternvoidspace, that's the same - 172.16.0.x masked with 255.255.255.0 yields 172.16.0.015:18
voidspaceyep, the netmask is the same15:18
voidspaceso they are actually the same15:18
voidspaceI just have to construct the cidr correctly15:18
voidspaceok15:18
voidspacedimitern: thanks15:18
voidspacedimitern: so, we could still take an instance id and only pair up networks from the filtered list with nodegroup interfaces15:19
voidspaceas we'll still need the call to the networks api15:19
dimiternvoidspace, let me think for a bit..15:19
dimiternvoidspace, consider having *a lot* of networks, but a given node only uses a few15:20
voidspacethen having a filtered list is useful15:20
dimiternvoidspace, it seems reasonable to filter by instance id, if possible to only get the relevant networks15:20
voidspaceyep15:20
voidspacedimitern: cool, I'll undo that change then...15:21
dimiternvoidspace, cheers, and sorry for the round-tripping :15:21
dimitern:)15:21
voidspace:-)15:22
=== kwmonroe_ is now known as kwmonroe
alexisbcherylj, welcome!16:06
alexisball cherylj just joined the Juju Core team16:06
perrito666well wellcome to all cherylj16:07
jw4cherylj: welcome!!16:07
ericsnowcherylj: congrats and welcome!16:07
cheryljthanks everyone :)16:07
cheryljI'm super-psyched to be part of the team16:07
TheMuecherylj: heya and welcome to the team from Germany too16:13
* perrito666 got some browser to work properly after only 2 months16:14
cheryljthanks, TheMue16:14
alexisbI have to day it is oddly quiet in this channel today :)16:16
perrito666alexisb: it is a big machinery it takes a bit to get it rolling again16:16
alexisbperrito666, heh :)16:17
voidspaceg'night all16:21
voidspaceoff to see Phantom of the Opera16:21
=== kadams54 is now known as kadams54-away
=== urulama is now known as urulama|kids
=== kadams54-away is now known as kadams54
katcocherylj: welcome to the team, cherylj :)17:32
cheryljthanks, katco :)17:37
=== kadams54 is now known as kadams54-away
=== kadams54-away is now known as kadams54
jogcherylj, welcome!17:55
katcocherylj: btw if you have any questions about go or otherwise, feel free to ask here or ping me personally. i promised thumper i would help however i could! :)17:56
cheryljHey jog, long time no see ;)18:21
cheryljAnd thanks, katco.  I'm still making my way through the New Starter Tasks and discovering all the things that I still don't have access to18:22
katcocherylj: fun times :)18:22
perrito666cherylj: katco think of it as a videogame, then we can all get together at a sprint and remember: "how did you beat the level where the dragon blocks the pop3 mail access?"18:25
katcoperrito666: haha18:25
katcoperrito666: how many lives do we get, and are we on easy mode or hard mode?18:25
perrito666katco: good question, I don't intend to find it18:26
perrito666:p18:26
katcohehehe18:26
perrito666I still recall that level where my phone holding 2fa died on me 14k km away from home18:26
katcothat is the stuff of nightmares18:27
perrito666and I was using a brand new os, which held none of my login cookies18:28
* perrito666 lights a candle to IS18:28
katcolol18:28
sinzuikatco, perrito666 : I don't see many Core on line...do either of you have time to look into bug 140769918:29
mupBug #1407699: Precise services cannot be deployed <ci> <deploy> <precise> <regression> <juju-core:Triaged> <https://launchpad.net/bugs/1407699>18:29
sinzuikatco, perrito666 I need to capture some logs from a machine 1 or 2 to get more information18:30
perrito666mmm I have seen this before18:32
perrito666sinzui: Ill take it18:34
katcoi've seen it stuck in pending before, but i'm not sure if it's for the same reason18:34
perrito666I have seen the particular error in functional ha backup and restore18:35
perrito666it is a bit more detailed since it breaks while trying to fetch a given file18:35
perrito666sinzui: precise-dummy charm is publicly available?18:38
sinzuiperrito666, yes., but it happens with the precise ubuntu charm too18:38
perrito666sinzui: I am trying to make sure Ill try to correct incantation, Ill check the config for the test env18:39
sinzuiperrito666, https://code.launchpad.net/~juju-qa/juju-ci-tools/repository is the charm repo used in tests.18:39
* perrito666 pulls master for the second time this day and gets updates... isn't merge blocked?18:45
sinzuiperrito666, its a cloud-init issue18:46
sinzuiperrito666, I will attach the logs to the bug18:47
jogsinzui, I'm also seeing '/bin/sh: 1: exec: cloud-init: not found' at the console for the maas nodes that were deployed for precise charms18:51
sinzuiperrito666, https://bugs.launchpad.net/juju-core/+bug/1407699 looks like failure in packages18:51
mupBug #1407699: Precise services cannot be deployed <ci> <deploy> <precise> <regression> <juju-core:Triaged by hduran-8> <https://launchpad.net/bugs/1407699>18:51
sinzuijog I expect the maas tests to be using trusty as the tests claim. Are they?18:52
jogsinzui, bootstrap node is trusty but the dummy-source and dummy-sink are precise18:54
sinzuijog, that is odd, I know we can do that, but we usually make them match. Anyway. we know know all testfailures match for br 219518:54
=== kadams54 is now known as kadams54-away
=== urulama|kids is now known as urulama
=== kadams54 is now known as kadams54-away
=== kadams54-away is now known as kadams54
thumpercherylj: welcome19:58
cheryljthumper:  thanks!19:59
thumpercherylj: do you have hangouts set up?20:01
thumpercherylj: ISTR that you do20:01
=== kadams54 is now known as kadams54-away
=== kadams54-away is now known as kadams54
=== kadams54 is now known as kadams54-away
thumperperrito666: how goes the looking at the critical bug?20:44
thumperperrito666: it smells like a missing package...20:44
perrito666thumper: almost sure its a missing package20:49
perrito666trying to get to know a bit more of cloudinit to know what is going on20:49
thumperperrito666: have you poked the cloudinit folks?20:49
thumpersmoser I think20:49
perrito666sinzui: if this ends up being a "third party" issue, how is it handled? I am guessing it trascends just saying "not our problem", just open a bug to the right people?20:59
waiganimenn0: did you get hit by another quake? http://www.stuff.co.nz/national/64684438/large-earthquake-jolts-canterbury-awake21:06
menn0waigani: yep we did. it didn't feel too bad but was quite different to the other little jolts i've felt while i've been here21:07
sinzuiperrito666, use "Also affects project"21:07
menn0waigani: it went on for quite a while and left the lights in hallway swinging for ages21:07
sinzuiperrito666, also, why isn't older juju affected, or trusty?21:07
waiganimenn0: yikes, glad you're all okay21:08
thumpersinzui: older juju passes fine on precise?21:08
thumpersinzui: I have just been reading the cloud-init log21:09
sinzuithumper, yep, every upgrade test used 1.20.x to boostrap the same stack21:09
waiganimenn0: I now have an image of you unicycling in an earthquake21:09
thumpersinzui: hmm...21:09
thumpersinzui: can we get the actual cloud init file off the machines?21:10
thumpersinzui: and compare the 1.20 cloudinit file with the 1.22-alpha?21:10
sinzuithumper, the cloud-init-output.log?21:10
thumperno, the source21:10
* thumper looks for the location21:11
thumpersinzui: so.. bootstrapping works on precise but adding machines doesn't?21:12
thumpersinzui: is that right?21:12
sinzuithumper, yep21:13
thumpersinzui: ok, unlikely to be an actual cloud init problem then...21:14
=== kadams54 is now known as kadams54-away
sinzuithumper, the log I attached to the bug shows that cloud-init was removed during the upgrade and install steps...should it have been?21:16
thumperI don't know21:17
* thumper pokes locally21:17
thumpersinzui: local precise fails too right>?21:17
sinzuithumper, 1.20.14 and 1.21-beta4 can deploy precise services fine. I have a precise machine deployed by 1.21-beta4. What do you want?21:18
sinzuiwell I can get the cloud-init-output.log and see if cloud-init was removed21:18
thumpersinzui: I'm looking...21:18
thumpersinzui: that would be helpful21:18
=== kadams54-away is now known as kadams54
* perrito666 has a broken machine21:20
thumperhmm...21:22
thumpermy local juju deployed precise just fine using local21:23
thumpersinzui: looking at this failure: http://juju-ci.vapour.ws:8080/job/local-deploy-precise-amd64/2275/21:24
thumpersinzui: the cloud-init log isn't - it is a juju log file21:24
sinzuithumper, that isn't the bug :)21:25
sinzuithumper, I attached the cloud-init-output.log to the bug21:25
thumperhowever it is still an issue21:26
thumperI was looking at the local deploy failure because my one worked21:26
sinzuithedac, https://bugs.launchpad.net/juju-core/+bug/1407699/comments/221:26
mupBug #1407699: Precise services cannot be deployed <ci> <deploy> <precise> <regression> <juju-core:Triaged by hduran-8> <https://launchpad.net/bugs/1407699>21:26
perrito666:| ls: cannot access /usr/lib/python2.7/dist-packages/cloudinit/: No such file or directory21:27
sinzuiperrito666, right the log shows cloud-init was uninstalled21:28
thumperperrito666: it removed cloud init...21:28
thumperwhich seems weird21:28
thumperright?21:28
sinzuithumper, my deploy of precise/ubuntu with 1.21-beta4 is much shorter.  it doesn't go into a phase to remove anything21:28
menn0thumper: but your machine-0 is still trusty right? or do you have a precise machine/VM too?21:28
thumpermenn0: I deployed precise/ubuntu to a machine21:28
thumperto get a precise lxc started21:28
thumperand it worked fine21:29
thumperalthough, this is from a pre-christmas trunk21:29
* thumper switches to master and rebuilds21:29
menn0thumper: so we think it's a problem with any precise machine, even if other machines are on trusty?21:29
thumperI'm looking21:30
thumpersinzui: that cloud init file, was that from an lxc machine or a clout machine?21:30
perrito666I deployed a just compiled version and created a stream and deployed to aws and it broke straight forward21:30
sinzuithumper, machine 1 and machine 2 (both lxc in the case of local provider) had their cloud-init packages removed21:31
sinzuiI didn't post machine2 since the log is almost identical21:31
perrito666I found it21:32
thumperperrito666: yes?21:32
perrito666cloud-image-utils is being installed21:32
sinzuithumper, this looks the the key events. we see cloud-init being upgraded, some deferment, then something decides to do another round of upgrades that removes cloud-init http://pastebin.ubuntu.com/9678855/21:32
perrito666it replaces cloud-utils21:32
perrito666cloud-init deppends on cloud-utils21:32
perrito666and therefore gets whacked21:32
thumperand we are getting cloud-image-utils because?21:33
thumperI wonder if that was added recently21:33
perrito666thumper: seems to be part of an upgrade?21:34
perrito666cloud init gets upgraded21:35
thumperperrito666: I see in our code we explicitly install it21:36
perrito666thumper: oh?21:36
thumperwhy oh why doesn't git have a nice history viewer like bzr explorer21:36
thumperis there a nice git equivalent of 'bzr qannotate' ?21:37
perrito666it was part of Ians fix before the shutdown seems21:37
menn0yep, 15th December21:37
thumperperrito666: got a PR?21:37
perrito666https://github.com/juju/juju/commit/e6cf44d021:38
perrito666got a commit21:38
thumperperrito666: well, he is due to start his work year shortly, so I say we ask him "-)21:39
thumperwallyworld_: oi... you broke CI :-)21:39
thumperwallyworld_: welcome back to work21:39
wallyworld_thumper: hi, i have a branch which i proposed before the break which fixes it if it is the issue i am thinking of21:40
perrito666that sounds awfully like my car's mechanic: "oi, remember when you fixed the GNC injector, you broke the fuel pump"21:40
wallyworld_i am ina meeting, will talk to you soon21:40
menn0thumper, perrito666, wallyworld_: it's PR 120521:40
thumperwallyworld_: for reference, https://github.com/juju/juju/commit/e6cf44d0 introduced the package cloud-image-utils, which breaks precise deployments21:40
perrito666and as menn0 said, it is https://github.com/juju/juju/pull/120521:41
perrito666lemme make that into an assigned issue :)21:42
perrito666wallyworld_: left you a nice comment and assigned you to the bug, cheers21:45
perrito666happy new year :p21:45
wallyworld_perrito666: thanks, will look after meeting21:45
wallyworld_you too21:45
* perrito666 wonders why he set up is office in such a manner that he needs to pick one of Air Conditioning or Big screen for the computer21:52
thumpercherylj: got time for a chat?21:53
cheryljthumper:  yes, ready whenever you are21:53
wallyworld_thumper: just finished meeting, reading backscroll, cloud-image-utils is needed to pull in ubuntu-cloudimg-query22:30
thumperwallyworld_: otp22:31
wallyworld_ok22:31
perrito666wallyworld_: I am guessing this might be an issue on cloud-init package dependencies in precise?22:31
wallyworld_perrito666: yeah, sadly, seems like a precise only issue22:31
waiganidavecheney: creating a new watcher for each assert in the test works.22:32
wallyworld_perrito666: i'll have to figure out the magic sauce needed to make precise work22:32
waiganidavecheney: I had to set the revno of each new watcher to the revno of the last watcher, as it gets updated if there was a change22:32
alexisbwallyworld_, do you think we will need to make a request to the cloud-init team?22:33
alexisbif so I will give gaughen a heads up22:33
wallyworld_alexisb: hey. not sure yet, i'll do some more digging first22:33
perrito666wallyworld_: I haven't looked at the other distros but cloud-init should deppend on a virtual (i cant recall the actual name) package satisfied by both utils and image-utils22:33
alexisbwallyworld_, ack, o and I guess I should say welcome back :)22:34
wallyworld_yeah,you too22:34
waiganidavecheney: my implementation of updating the revno for each watcher feels ugly, so I might poke you once the branch is up for review to see if there is a neater way22:34
=== kadams54 is now known as kadams54-away
=== kadams54-away is now known as kadams54
perrito666wallyworld_: ok a bit of extra info, in utopic it deppends on cloud-utils or cloud-guest-utils22:35
perrito666guest-utils is pulled by cloud-image-utils22:35
perrito666thus satisfying the dependency22:36
wallyworld_perrito666: when you run ubuntu-cloudimg-query, it twlls you to install cloud-image-utils, so that's why i added that dependency22:36
perrito666wallyworld_: this seems to be only broken on precise22:36
wallyworld_yes, that is my fear22:37
perrito666you need to have cloud-init 0.7.6~bzr1022-0ubuntu1 available it would seem22:37
perrito666to met the dependency22:37
perrito666given the metadata on all three packages you need to get a hold of scott moser22:38
wallyworld_i recall cloud-init being updated recently22:38
alexisbwallyworld_, smoser is on #server22:39
wallyworld_ok, thanks22:40
=== kadams54 is now known as kadams54-away
=== kadams54-away is now known as kadams54
perrito666ok, eod people, see you all tomorrow22:59
menn0thumper: can you PTAL at this PR from yesterday? http://reviews.vapour.ws/r/672/23:23
=== kadams54 is now known as kadams54-away
menn0thumper: also, you were asking about a git equivalent to bzr qannotate earlier23:27
menn0thumper: if your editor doesn't support such a think then have a look at "git gui blame <path>"23:27
menn0s/think/thing/23:28
thumpermenn0: does that allow you to walk back through history?23:33
menn0thumper: I believe so. you can click on a rev for a line. I haven't used it much.23:33
thumpermenn0: I don't think I'm using any sublime git plugins23:35

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!