/srv/irclogs.ubuntu.com/2012/03/05/#juju.txt

=== al-maisan is now known as almaisan-away
=== almaisan-away is now known as al-maisan
=== al-maisan is now known as almaisan-away
=== almaisan-away is now known as al-maisan
=== al-maisan is now known as almaisan-away
jcastronegronjl: mira, the moin charm got some updates a while back but we never finished it off, it looks pretty close if you want to finish off a review15:49
negronjljcastro: I'll look into it in a minute....id you happen to have a bug number, that will speed things up16:41
negronjlSpamapS: I updated distcc ( https://bugs.launchpad.net/charms/+bug/944989 ).  When you get a chance, can you review it again ?16:41
_mup_Bug #944989: charm needed: distcc <Juju Charms Collection:In Progress by negronjl> < https://launchpad.net/bugs/944989 >16:41
SpamapSnegronjl: looking righ tnow16:47
negronjlSpamapS: thx.16:47
SpamapSnegronjl: *really* nice work16:52
negronjlSpamapS: thx ... ok to promulgate now ?16:53
SpamapSnegronjl: do you want to promulgate that? (don't forget to make it owned by ~charmers first)16:53
negronjlSpamapS: Yes and yes ...16:53
SpamapSI'll +1 in the bug just for doc purposes16:53
negronjlSpamapS: thx16:53
negronjlSpamapS: btw ... I thought that promulgate would change the ownership of the branch to ~charmers on it's own ...16:54
SpamapSnegronjl: it doesn't16:54
negronjlSpamapS: ok .. .I'll do it16:54
SpamapSnegronjl: I believe pushing to lp:charms/foo does ..16:55
negronjlSpamapS: We'll find out very soon :)16:55
SpamapSif that doesn't do it, then I've taken to just pushing to lp:~charmers/charms/oneiric/foo/trunk with --rememeber .. then promulgating16:55
negronjlSpamapS: distcc ... promulgated17:09
negronjlglusterfs charm .... here I come !!!!!!!!17:09
SpamapSoooohhhhhh snap17:12
m_3SpamapS: hey, can you please change ownership of lp:charms/tomcat6 to ~charmers?17:13
negronjlSpamapS: will do17:13
m_3negronjl: can you do the same for lp:charms/hadoop-mapreduce?  that's still showing as a personal branch17:14
negronjlSpamapS: lp:charms/tomcat6 shows as yours :P17:16
negronjlm_3: will do17:16
negronjlm_3: done17:17
m_3negronjl: cool thanks!17:17
SpamapSnegronjl: doh17:17
negronjlm_3: no worries ...17:17
jamespagem_3: zookeeper charm appears to work OK on oneiric17:18
m_3jamespage: whoohoo!17:18
jamespagem_3: should I remove the config options that are specific to precise?17:18
jamespageits a bit misleading ATM17:18
m_3jamespage: I think they'll be two separate charms maintained in parallel... any better ideas?17:19
jamespagem_3: kinda what I mean; I'll update the branch to remove the PPA options for oneiric17:19
jamespagethey can stay for the precise version17:19
m_3jamespage: right17:19
jamespagem_3: or I could backport the packages in the PPA to oneiric - that would be easier actually17:20
jamespagethen people can kill themselves with 3.4.3!17:20
m_3jamespage: easier?  really?  :)17:20
jamespagem_3: the packaging is simple so yes17:21
m_3jamespage: it makes sense to keep separate branches lp:~charmers/charms/oneiric/zookeeper/trunk (lp:charms/zookeeper) and lp:~cha...precise...17:21
jamespagem_3: yep17:22
m_3jamespage: it's just a little backwards from the usual "direction" of the branch :)17:22
jamespagem_3: charm backport - I like it!17:22
* m_3 groan17:22
jamespagem_3: OK if I promulgate that now!17:23
jamespage?17:23
m_3jamespage: just still scared of backports in packages :)17:23
m_3jamespage: please go ahead and shrink wrap it17:23
jamespagem_3: https://code.launchpad.net/~charmers/charms/oneiric/zookeeper/trunk17:24
jamespagedeone17:24
jamespagedone even17:24
m_3jamespage: awesome!17:25
m_3charms are almost tended... jcastro will even be happy... we're net positive count after the changes :)17:25
jcastro\o/17:26
andrewsmedinaWhat I do when a bootstrap instance state is not-started?18:40
avoinecheck the log on the server in /var/log/juju/18:41
andrewsmedinaavoine: -bash: cd: /var/log/juju: No such file or directory18:43
andrewsmedina=/18:43
avoineif there is none and you are on aws also check /var/log/cloudinit/18:43
SpamapSandrewsmedina: what provider are you using ?18:44
andrewsmedinaSpamapS: ec2 (openstack)18:44
andrewsmedinausing default juju version works fine, but with the last ppa version, bootstrap returns "not-started" status18:45
SpamapSbootstrap usually doesn't return anything.. can you explain?18:45
andrewsmedinaSpamapS: after bootstrap, when I execute juju status, returns the bootstrap instance status18:46
avoineandrewsmedina: check in /var/log/cloudinit/ if something have failed there18:49
SpamapSandrewsmedina: right, ok. So you need to use the EC2 API tools to figure out why the instance hasn't started yet.18:50
SpamapSandrewsmedina: euca-get-console-output on the instance ID should help18:50
hazmatits probably the same problem of the metadata server not being configured to return values for  1.0 api version18:51
hazmatits a configuration (.ini) fix for nova18:51
hazmatie. bug 94286818:52
_mup_Bug #942868: metadata service reports to support 1.0 but serves incorrect data <canonistack> <juju:Confirmed> <OpenStack Compute (nova):New> < https://launchpad.net/bugs/942868 >18:52
SpamapSRight!18:53
hazmatit feels like a packaging bug, we could update juju to just use latest or another common version that's enabled by default18:53
SpamapSmajor DOH there ;)18:53
hazmatie. work around it18:53
andrewsmedinaSpamapS: is everything ok in the instance18:53
* SpamapS is getting giddy.. in 2 hours his upstream will go from 1Mbit to 5Mbit .. woot.18:53
SpamapShazmat: I think we should update juju to use the *latest* common version that we know we can support.18:54
hazmatSpamapS, technically the metadata we're after is common to all of the extant versions18:54
andrewsmedinaSpamapS: but I think that juju does not found zookeeper agent in this instance18:55
SpamapShazmat: isn't there a "best effort" url?18:55
hazmatit looks like 2009-04-04 is the latest version available18:55
andrewsmedinaSpamapS: cloud-init boot finished at Mon, 05 Mar 2012 18:53:55 +0000. Up 300.53 seconds18:55
hazmatSpamapS, 'latest' is latest, but some risk of deviation..18:55
hazmatpretty minor though18:55
hazmatinternally these all dispatch to the same implementation for ostack18:55
SpamapSandrewsmedina: right, so from there, you need to look at the juju agent logs,which will be in /var/log/juju .. or perhaps if they had catastrophic errors, in /var/log/upstart18:56
hazmatthe error should be in the console output18:56
SpamapShazmat: try:;latest;except:;try:known_old_version18:56
jcastrohazmat: ok so now /charms says 60 official18:56
hazmatjcastro, woot!18:57
hazmatm_3, nice18:57
jcastrohazmat: heh ok, so we're sure this is right right?18:57
andrewsmedinaSpamapS: but with an old juju version works fine18:57
jcastroI still see some are counted twice18:57
hazmatjcastro, /charms is cross series, add a series to it if you want a particular series18:58
hazmatjcastro, ala http://charms.kapilt.com/charms/oneiric18:58
jcastrooh cool!18:58
jcastrogot it!18:58
jcastrook, so 54!18:58
SpamapShazmat: with the agents being started in upstart jobs.. are you sure the error would be in the console?18:58
hazmatSpamapS, yes, the metadata server error happens pre-agent during juju-admin initialize18:59
hazmatwhich is just a cli command for cloud-init18:59
hazmatandrewsmedina, can you pastebin the console output?19:00
SpamapShazmat: well at least we have a fail early then. ;)19:00
andrewsmedinaSpamapS: http://dpaste.org/u7oDI/19:06
SpamapShazmat: ^^19:08
hazmatandrewsmedina, that's status output.. i meant the equivalent of ec2-console-output.. or the cloud init log19:09
hazmatbut that's the output i would expect for this bug from status19:11
andrewsmedinahazmat: http://dpaste.org/1qhg4/19:11
hazmathuh..19:11
andrewsmedinaeverything seems ok19:12
hazmatandrewsmedina, if your logged on the machine.. what's the output of.. curl http://169.254.169.254/1.0/meta-data/instance-id19:12
andrewsmedinahazmat: humm19:14
andrewsmedinahazmat: http://dpaste.org/oRDnk/19:14
andrewsmedinahazmat: juju-sample groups isn't created19:17
andrewsmedinahazmat: im wrong19:19
andrewsmedinajuju-sample groups is here19:19
andrewsmedinahazmat: the problem is my http_proxy =/19:25
=== almaisan-away is now known as al-maisan
=== al-maisan is now known as almaisan-away
SpamapSUgh always the proxy19:39
hazmaterrand time, bbiab19:47
andrewsmedinaI removed the proxy but nothing has changed =/20:00
jcastronegronjl: nice work on the review!20:21
jcastroso we can promulgate that badboy?20:21
negronjljcastro: sure20:21
negronjljcastro: checking to see if he has access to do it.  If he doesn't, I'll merge it.20:22
jcastroI don't think he does20:22
jcastrohe looks like a new contributor20:23
jcastrotime for a tshirt and a mug!20:23
negronjljcastro:  python-moinmoin merged and pushed and promulgated20:26
* jcastro rings the bell20:26
negronjlclear20:26
jcastrothird one in 2 days fellas!20:26
negronjljcastro:  do we need to upgrade this guy ?20:26
jcastrowhat do you mean?20:27
negronjljcastro: how many charms does patrick have ?20:27
jcastrothis is his first20:27
negronjljcastro: nm then ... ignore me20:27
jcastroyour mixing him up with Nathan I think20:28
jcastrowhich has like 4 or 5 now or something20:28
jcastrobut only interested in writing, not reviewing20:28
jcastroSpamapS: got time for a quick G+?20:44
jcastroI need like 45 seconds for you to explain something to me20:44
jcastrom_3: https://launchpad.net/~chris-lea/+archive/node.js20:53
jcastroFYI I didn't know there was a node PPA20:53
_mup_Bug #947483 was filed: Support for container relations <juju:New> < https://launchpad.net/bugs/947483 >20:59
hazmatbcsaller, lbox needs the bug flag to show up in the kanban21:00
bcsallerhazmat: and I think rerunning it isn't adding the bug for me now, but I think it used to21:01
hazmatbcsaller, i just did it by hand..21:02
_mup_juju/force-upgrade r458 committed by kapil.thangavelu@canonical.com21:08
_mup_some environment destruction tests where missing inlinecallbacks decorator21:08
hazmatniemeyer, the charm test spec looks like it could use another review.. https://codereview.appspot.com/5624044/21:19
hazmatm_3, what's the url for your charm testing jenkins instance?21:21
* hazmat keeps forgetting21:22
SpamapSjcastro: back from lunch, we can G+ any time.21:33
m_3hazmat: it's moving ec2 accts atm... gimme a few21:42
jimbakerhazmat, can you respond to julian edwards and gavin panella re the maas provider branch?21:54
* SpamapS still wishes providers had been made fully pluggable so they would be useful without being fully integrated into juju.environment.config :-P21:56
hazmatjimbaker, sure21:57
jimbakerhazmat, thanks!21:57
andrewsmedinahazmat: my problem is that juju version in bootstrap in different from my machine22:00
andrewsmedinahazmat: in my machine I use ppa version22:01
hazmatandrewsmedina, juju should detect that and use the ppa version when bootstrapping, unless a different juju-origin has been set in environments.yaml22:01
jimbakerSpamapS, good point. a role for ju-jujitsu? or is that just crazy talk?22:03
andrewsmedinahazmat: I defined juju-origin: ppa, but the version installed is the distro version22:04
SpamapSjimbaker: totally crazy. juju has to adapt to make them pluggable22:04
hazmatSpamapS, after the cli plugins got nixed i gave up on out of tree plugins22:05
hazmatSpamapS, its moot at this point... and in the future juju will be a binary cmd22:05
hazmatbig fat juicy binary22:05
jimbakerhazmat, :)22:06
SpamapScan still do plugins22:06
hazmatSpamapS, for providers.. maybe.. but they'd be out of process.. i wouldn't hold your breath22:06
SpamapShazmat: go has no dlopen?22:11
hazmatSpamapS, yeah.. it has LoadLibrary..22:14
jseutterhey, I'd like to test out my charm using lxc containers I saw mentioned a while back.  Does this work, and is there some documentation on how to do it?23:24
m_3SpamapS: where's the latest charm testing spec branch?23:36
m_3that moved into juju itself right?  or is it still in charm-tools?23:36

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