[05:32] <melmoth> is there a known issue with juju 1.12 with destroying-service ? Since i switch to juju >=1, destroy-service fails silently most of the time
[05:32] <melmoth> the service are shown as dying, but nothing else seems to happen
[05:41] <raywang> melmoth, i do have seen this
[05:42] <melmoth> i do quite a lot those days
[05:42] <raywang> melmoth, me too
[12:24] <bac> hi mectors
[13:29] <juju> hi
[13:30] <juju>  8-)
[13:30] <juju>  :p
[13:34] <juju> hi
[13:35] <juju>  hi is any one her
[14:01] <geme> jcastro, can Debian / Redhat images be deployed ?
[14:02] <jcastro> not currently no
[14:03] <jcastro> I think debian has cloud-init now, all that's needed is for someone to do the work
[14:04] <geme> I tried using Debian with juju awhile back - problems with cloud-init confg & upstart
[14:05] <hazmat> jcastro, re the rack charm, tell the author he has bug in his config.
[14:06] <hazmat> that's preventing the store publishing, fwiw
[14:06] <geme> jcastro, Can different ubuntu images be used in the same environment ?
[14:09] <jcastro> geme: like different versions?
[14:09] <jcastro> or different images as in cloud images?
[14:09] <jcastro> hazmat: got any more detail?
[14:10] <jcastro> hazmat: or is it that encoding issue sinzui and him ran into?
[14:10] <geme> yes - different ubuntu cloud images ?
[14:11] <jcastro> we do what is defined in "series" in environments.yaml, we don't have a way to specifify a custom image, for example
[14:13] <hazmat> jcastro, nothing to do with encoding he has a bad config default value .. config: options.deploy_key.default: unexpected value <nil>
[14:13] <geme> So, just the one image available for say precise ?
[14:13] <jcastro> geme: yeah it will fetch the latest supported cloud-image and deploy that
[14:14] <jcastro> http://cloud-images.ubuntu.com/releases/precise/release/
[14:30] <geme> jcastro, is there a timescale to support other oses ?
[14:35] <jcastro> not really, any help would be accepted and appreciated, but we're not actively working on that at the moment
[15:28] <arosales> smoser, utlemming: talking with mgz on bug https://bugs.launchpad.net/juju-core/+bug/1188126
[15:28] <_mup_> Bug #1188126: Juju unable to interact consistently with an openstack deployment where tenant has multiple networks configured <canonistack> <openstack> <serverstack> <juju:New> <juju-core:Triaged> <https://launchpad.net/bugs/1188126>
[15:28] <arosales> any way to tell cloud-init to bring up all detected networks?
[15:32] <utlemming> arosales: generally with a cloud-drive you can send /etc/network/interfaces, but if they are using the EC2 meta-data source, then no
[15:34] <arosales> utlemming, (brainstrom) any way to patch cloud init to take ingest a value to detect and init all attached networks (ie not stop in the first)
[15:39] <arosales> smoser, ^
[15:48] <hazmat> arosales,  i don't think it an issue of bringing up the network
[15:49] <hazmat> afaik, the issue is that is being able to correctly identify which network juju should be using
[15:49] <hazmat> given multiple
[15:49] <arosales> hazmat, just an idea mgz and I were batting around. mgz would have more insights then myself
[15:50] <geme> jcastro, attempting 1st bootstrap using juju-core 1.13.1 and it errors - http://*.*.*.*:8080/v2.0/AUTH_ca452007af584a7f933e6a3e281f1c85/juju-dist/"/"streams/v1/index.sjson": cannot find URL "http://*.*.*.*:8080/v2.0/AUTH_ca452007af584a7f933e6a3e281f1c85/juju-dist/streams/v1/index.sjson" not found
[15:50]  * hazmat drops back to #juju-dev for it
[15:50] <mgz> hazmat: I don't see why, in the elmo-case, we shouldn't expect both to work
[15:50] <geme> should index.sjson be index.json
[15:50] <mgz> (ie, they're trying to migrate an existing deployment to a new ip range, by adding an extra address to every instance)
[15:51] <marcoceppi> geme: does it actually fail? AFAIK that's a common error during juju bootstrap -v but doesn't nessiarily mean it's failing. Is this a private openstack deployment?
[15:52] <geme> yes - it's a private openstack
[15:53] <marcoceppi> geme: Do you see any instances online started by juju in the dashboard? does juju status work? Was that the error that juju threw during bootstrap (actually, could you pastebin the the full output to the shell)
[15:53] <geme> No instance deployed
[15:53] <arosales> geme, sounds similar to the https://lists.ubuntu.com/archives/juju/2013-August/002826.html thread
[15:54] <marcoceppi> Have you created image metadata for your dpeloyment using the juju metadata command?
[15:54] <marcoceppi> finally, what version of juju are you using? (juju version)
[15:54] <geme> yes - image metadata created and posted to the public-bucket
[15:55] <geme> juju-core 1.13.1
[15:55] <geme> error: cannot start bootstrap instance: no "precise" images in RegionOne with arches [amd64 i386]
[15:56] <geme> pastbin url ?
[15:57] <geme> pastebin url
[15:57] <arosales> geme, http://pastebin.ubuntu.com/
[15:57] <marcoceppi> geme: http://paste.ubuntu.com
[15:57] <marcoceppi> ^^
[15:58] <smoser> arosales, sorry. back now.
[15:58] <smoser> ah. for ec2, there is a well identified solution.
[15:58] <marcoceppi> geme: juju will attempt to access several locations. starting with .sjson, then .json, then public simplestream data
[15:58] <smoser> that is simply just not implemented.
[15:58] <smoser> https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/1153626
[15:58] <_mup_> Bug #1153626: Multiple Interfaces and IPs not detected in AWS VPC <aws> <cloud-images> <ec2> <vpc> <cloud-init:Triaged> <cloud-init (Ubuntu):Triaged> <https://launchpad.net/bugs/1153626>
[15:59] <smoser> arosales, ^
[16:02] <geme> marcoceppi, bootstrap -vv console output in pastebin
[16:03] <marcoceppi> geme: can you paste the link please?
[16:03] <geme> http://paste.ubuntu.com/5989532/
[16:19] <marcoceppi> geme: what's the region set to in your index.json file?
[16:20] <geme> RegionOne
[16:20] <marcoceppi> geme: could you paste the results of that file in a pastebin
[16:21] <marcoceppi> also, it's nearing bed time for me. So if you're unable to get futher assistance, feel free to post to ask ubuntu or juju@lists.ubuntu.com for "offline" help
[16:21] <geme> index.json ?
[16:21] <marcoceppi> geme: yes
[16:21] <marcoceppi> geme: I'm also assuming the *.* in the addresses are just you anonymizing the data?
[16:21] <geme> yes
[16:23] <geme> http://paste.ubuntu.com/5989600/
[16:24] <marcoceppi> geme: you dont' actually have any cloud data in there
[16:26] <geme> ran this command - juju-metadata generate-image -i 65a1e7ba-eb56-48e6-9b4e-8f7c0c05e779  -s precise -r RegionOne -u http://15.25.24.11:5000/v2.0/
[16:26] <geme> Am I missing something ?
[16:26] <marcoceppi> geme: not sure he's an example working index.json file for hpcloud
[16:27] <arosales> smoser, the conversation continued in #juju-dev between mgz and hazmat
[16:28] <geme> marcoceppi, so the index.json is incorrect ?
[16:29] <marcoceppi> geme: it's missing data, one sec
[16:47] <marcoceppi> geme: what does imagemetadata.json look like?
[16:48] <geme> http://paste.ubuntu.com/5989688/
[16:49] <geme> marcoceppi, need to drop off. Pick up tomorrow ?
[16:49] <marcoceppi> geme: yes, I need to do the same
[16:50] <geme> ok - catch you tomorrow. Thanks
[18:17] <sidnei> uhm, where do i make changes to https://juju.ubuntu.com/docs/charms-constraints.html ?
[18:22] <jcastro> arosales: wotcha think: https://juju.ubuntu.com/events/
[19:11] <juju> hi-
[19:37] <arosales> jcastro, looks good, thanks for updating it
[19:37] <arosales> lot of topics I am really interested in there too
[19:37] <arosales> good content
[20:43] <juju> hi
[21:00] <marcoceppi> sidnei: lp:juju-core/docs
[21:01] <sidnei> thanks marcoceppi
[23:31] <irossi> jamespage, How's it going? Was wondering if you could help me with some strange issues?
[23:32] <irossi> Hey can someone help me troubleshoot a strange Juju issue?
[23:36] <arosales> irossi, its end of day for europe and america time zones so may be good to mail the list and get some help there.
[23:36] <arosales> https://lists.ubuntu.com/mailman/listinfo/juju
[23:36] <arosales> or leave a post on askubuntu
[23:36] <irossi> OK thanks very much
[23:36] <arosales> irossi, I can try to help, but may have to point you to the list or askubuntu
[23:38] <irossi> arosales, Thanks. I have Openstack on 28 nodes deployed by Juju. Every once in a awhile, the juju machine agent starts to hog a lot of memory and I can't access the nodes.
[23:38] <arosales> irossi, which juju version are you running?
[23:38] <irossi> arosales, I'm able to reboot the bootstrap node, but it still shows instance-state: unknown while the agent-state is running
[23:39] <sarnold> irossi: check log sizes, I think I've heard that juju logs never get rotated and can fill the disk they are on..
[23:39] <thumper> we should fix that...
[23:39] <sarnold> irossi: check df output on the inolved machines..
[23:39] <arosales> sarnold, ah good point
[23:39] <irossi> arosales, sarnold Checking on all of the above
[23:39] <sarnold> thumper: we may have already, this might be old cargo-culting :)
[23:39]  * thumper nods
[23:40] <sarnold> all I know is tracking down problems due to full filesystems is -always- harder than it sounds.
[23:40] <sarnold> so check them first :)
[23:40] <arosales> sarnold, thumper so juju rotates logs on 1.13 >
[23:40] <irossi> Most of the nodes are so pegged I can't ssh into them
[23:41] <irossi> arosales, sarnold: Juju 0.7
[23:41] <sarnold> arosales: excellent! thanks. :)
[23:42] <arosales> sarnold, I was actually asking :-)
[23:42] <arosales> I should have appended a "?" to that
[23:42] <irossi> arosales, sarnold: Checking on the logs too
[23:44] <sarnold> arosales: oh :) haha
[23:44] <arosales> sarnold, but thumper would know the latest on log rotate in 1.13
[23:44] <irossi> arosales, sarnold : So it Juju 0.7 too old? Are there bugs in that version that are good reason to upgrade? What's the recommended version?
[23:45] <irossi> arosales, sarnold : 1.13 ?
[23:45] <arosales> irossi, I don't know for certain the issue you are hitting is addressed in 1.13, but I would recommend it
[23:46] <thumper> arosales: I don't actually know about the log rotation...
[23:47] <irossi> arosales, What's the easiest way to upgrade?
[23:47] <arosales> sudo add-apt-repository ppa:juju/devel && sudo apt-get update && sudo apt-get install juju-core gets you the latest
[23:47] <irossi> arosales, Why devel? That doesn't sound stable...
[23:47] <arosales> there is also sudo add-apt-repository ppa:juju/stable
[23:47] <arosales> which is at 1.12
[23:47] <thumper> haha
[23:47] <arosales> just was saying that devel got you the latest version if you were looking for a comparison
[23:49] <arosales> irossi, right now there is not a clean upgrade path from .7 to 1.13, but juju core devs are working on improving that.
[23:49] <arosales> irossi, probably not the answer you wanted to hear :-/
[23:49] <irossi> arosales, OK
[23:51] <arosales> irossi, juju core devs are working on bugs such as https://bugs.launchpad.net/ubuntu/+source/juju-core/+bug/1200878
[23:51] <_mup_> Bug #1200878: Upgrade breaks existing pyjuju deployment <apport-collected> <papercut> <regression-release> <saucy> <juju-core:Triaged> <juju-core (Ubuntu):Triaged> <https://launchpad.net/bugs/1200878>
[23:51] <arosales> and making a better transition story from .7 to 1.x
[23:52] <irossi> arosales, So the upgrade path you describe, would you mind describing it a bit more and perhaps any risk?
[23:55] <arosales> irossi, I would have to ping hazmat on the current upgrade story from .7 to 1.x
[23:55] <arosales> not sure if he around atm
[23:58] <irossi> arosales, Ok thx
[23:58] <irossi> arosales, I'll check in tomorrow
[23:59] <arosales> irossi, apologies, wish I had a better answer
[23:59] <arosales> irossi, but I also don't want to give you false info either.