[01:21] guys, is it possible to use Juju with Rackspace? [01:23] jose: yes [01:23] davecheney: where should I put the info? on the openstack part of the config [01:23] ? [01:24] depends, if rackspace looks close enough to openstack then you could try the openstack provider [01:24] or you could use the ssh provider and manually provision some machines for your environment [01:24] oh, I meant to use Rackspace automagically like it's done with AWS or HPCloud, for example [01:38] jose: if rackspace offers the same api endpoints as the openstack provider wants, then yes [01:38] jose: if it doesn't, then no [01:38] for using the openstack provider [01:38] I'm not exactly sure, that's why I was asking :) [01:38] I'll check, though [01:38] I'm not sure if any of the core devs have tried [01:38] I certainly haven't [01:47] jose: you need to use the manual provider. Rackspace doesn't have an object store. [01:47] ack, thanks [01:50] lazyPower: ah.. if that is all, hopefully soonish... we will be able to [01:51] thumper: thats teh only thing off the top of my head - i dont know about their API endpoitns [02:15] thumper: lazyPower for the longest time they didn't have cloud-init enabled on the OS images and such. [02:15] thumper: lazyPower that's changed in the last few months [02:16] jose: but I use the manual provide and works well if a bit more manual than hoped [02:16] cool, thanks rick_h_ [03:45] MAAS and Openstack Network | http://askubuntu.com/q/482090 [04:11] Add Nodes to MAAS for JUJU Bundle | http://askubuntu.com/q/482094 [04:49] Where to see the complete list of juju charms available? | http://askubuntu.com/q/482103 === vladk|offline is now known as vladk === CyberJacob|Away is now known as CyberJacob === CyberJacob is now known as CyberJacob|Away [11:06] is there a "preferred" way of including trademark/copyright information in a charm? specifically with regards the charm logo. [11:24] jamespage: ping. Can we talk about this Trusty SRU when you have some time please? [11:27] rbasak, sure - let me just eat lunch and I'll be back [11:27] OK [12:11] hi, while I thought I managed to fix this... I'm still having issues with the local provider: when I deploy some instances (eg, wordpress/mysql) the are kept in the pending state for a long time, and then they failed due to the template taking too long and then the second machine failing bc it couldn't clone a running template [12:12] I managed to work around this by destroying the environment and starting over (now that the template is finally complete) [12:12] but I thought it worth to mention [12:24] rbasak, back [12:24] o/ [12:25] jamespage: so I need to land this 1.18.4, which has ~20 bugs fixed. Trusty is on 1.18.0 at the moment. [12:25] This is something that I'd want an MRE for. Preparing SRU paperwork for 20 bugs is going to be a pain otherwise. [12:25] Is that what I need to do? I'm not sure all of them are individually SRU-worthy either. [12:26] Or, any other guidance on how to proceed? [12:26] Oh Trusty is on 1.18.1 I think, sorry. [12:26] Still, it's a fairly big bump in terms of commits. [12:28] rbasak, I agree that we need a MRE but this was pretty much nacked by pitti when I applied for it [12:28] rbasak, I'd email off that thread about what we need todo now and ask again [12:28] we can have a list of things that have been fixed [12:28] * rbasak looks [12:29] Ah, I see it [12:29] https://lists.ubuntu.com/archives/technical-board/2014-April/001877.html [12:31] jamespage: OK, thanks. I'll prepare the tracker bug and reply to https://lists.ubuntu.com/archives/technical-board/2014-April/001924.html I think. === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha [14:06] sparkiegeek: wrt you copyright/trademark question - yes. You can place that info in the Copyright file [14:06] there's a specific directive that marcoceppi showcased yesterday in his charm development track that shows how he denotes multiple levels of copyrighted material being included in a charm. [14:07] lazyPower: ok, great. Thanks === revagomes_ is now known as revagomes [14:56] arosales: yeah === isviridov is now known as isviridov|away [16:59] hey === psivaa is now known as psivaa-afk [16:59] is there any way to pass data between instances but not by relation-set? [17:05] Egoist: no [17:07] marcoceppi: ok. But data is taken from config.yaml, is ther any way to change this file while charm is running? [17:08] Egoist: no, you can only change configuration values from the user [17:09] charms can't change their own or others [17:10] I tried to follow http://askubuntu.com/questions/65359/how-do-i-configure-juju-for-local-usage and i got "error bootstrapping a local environment must not be done as root" [17:10] ali1234: which version of juju are you using? [17:11] the version in 14.04 [17:11] ali1234: that looks like your Ubuntu series. can you show me the output of juju --version [17:11] yes, it is my ubuntu series. i am using the version which ships with ubuntu 14.04 [17:12] 1.18.1-trusty-amd64 [17:12] do you have the juju-local package installed? it should have prompted you for sudo rights during the bootstrap [17:12] yes [17:12] i followed the instructions on askubuntu exactly [17:13] it did prompt for sudo password [17:14] http://paste.ubuntu.com/7634650/ [17:15] when i try to run it without sudo like it tells me to, it says this: http://paste.ubuntu.com/7634653/ [17:15] yikes [17:16] check dmesg output for anything 'strange'. that's mighty odd looking... [17:16] http://paste.ubuntu.com/7634668/ [17:17] all looks normal there [17:30] I'm trying to setup a juju local provider on my laptop here but can't seem to get it working. juju init works fine, I edit the config and change default to local, then juju bootstrap..... [17:30] juju status gives me this: [17:30] ERROR state/api: websocket.Dial wss://10.0.2.1:17070/: dial tcp 10.0.2.1:17070: connection refused [17:30] mongodb is in fact running on 37017 [17:30] there's no iptables rules in play [17:31] it doesn't seem like juju is creating the lxc container when bootstrapping [17:31] also seeing this in debug: [17:31] 2014-06-12 15:56:04 DEBUG juju.environs.simplestreams simplestreams.go:388 fetchData failed for "file:///home/leftyfb/.juju/local/storage/tools/streams/v1/index.sjson": stat /home/leftyfb/.juju/local/storage/tools/streams/v1/index.sjson: no such file or directory [17:35] marcoceppi: change configuration values from the user -> What does it mean? [17:35] also seeing this in /var/log/machine-0.log [17:35] 2014-06-12 16:54:39 DEBUG juju.agent agent.go:384 read agent config, format "1.18" [17:35] 2014-06-12 16:54:39 ERROR juju.cmd supercommand.go:305 symlink /home/leftyfb/.juju/local/tools/machine-0/jujud /usr/local/bin/juju-run: no such file or directory [17:37] Egoist: meaning that only users can change the values defined by config.yaml [17:37] lazyPower: Ok i get it. Thanks [17:38] leftyfb: looks like you're missing the simpestreams stuff that bootstrap should be fetching for you. is there a cached cloud image in /var/cache/lxc for cloud-trusty / cloud-precise? [17:38] ali1234: i'm sorry i dont know why its throwing a stack trace at you. thats a great candidate for a bug http://launchpad.net/juju-local [17:39] lazyPower: /var/cache/lxc/trusty/rootfs-amd64 exists with what looks like a complete filesystem [17:39] leftyfb: it should be creating cloud-trusty and cloud-precise iirc [17:40] the cloud-* are the cloud images that juju uses to create your containers. It may be prudent to 'nuke from orbit' your in progress components to build the lxc containers and start over. There's a plugin at http://github.com/juju/juju-plugins that warehouse this utility and has installation instructions - its 'juju clean' [17:41] I've got to get ready for my UOS track however, good luck. I'll stop in to check on you once my session is over [17:47] launchpad is kaput :( [17:48] lazyPower: that script/plugin only does 3 steps(and not sufficiently) that i've been doing over and over again trying to debug this issue. [17:50] ali1234: sorry, lazyPower had the wrong url, use this one: https://bugs.launchpad.net/ubuntu/+source/juju-core/+filebug [17:50] leftyfb: https://github.com/marcoceppi/plugins/commit/1c315f40fcc6e3102b4be5a4aa11544218159fd0 [17:50] sarnold: i did not use the url, i used ubuntu-bug [17:50] launchpad timed out while i tried to submit the report [17:50] ali1234: eep. [17:51] lazyPower: can you review/merge that? [17:51] i will just do it again [17:51] marcoceppi: i'd love to but i cant merge, i can only comment. [17:51] what, why not? Aren't you in the juju group? [17:52] WAIT [17:52] i was looking at your repo, whoops [17:52] w/e cowboy'd it [17:52] neat [17:53] cool, now i crashed apport [17:56] lazyPower: is this a typo somewhere? : [17:56] 2014-06-12 17:55:49 DEBUG juju.environs.simplestreams simplestreams.go:362 cannot load index "file:///home/leftyfb/.juju/local/storage/tools/streams/v1/index.sjson": invalid URL "file:///home/leftyfb/.juju/local/storage/tools/streams/v1/index.sjson" not found [17:56] should it be index.json , not index.sjson? [17:57] leftyfb: it checks for both [17:57] so it does [17:57] 2014-06-12 17:55:49 DEBUG juju.environs.simplestreams simplestreams.go:388 fetchData failed for "file:///home/leftyfb/.juju/local/storage/tools/streams/v1/index.json": stat /home/leftyfb/.juju/local/storage/tools/streams/v1/index.json: no such file or directory [17:57] #leftyfb@blanchard[0]:~$ ls -l .juju/local/storage/tools/streams/v1/index.json [17:57] -rw------- 1 leftyfb leftyfb 1152 Jun 12 13:55 .juju/local/storage/tools/streams/v1/index.json [17:58] nm, looks like it writes it further down [17:59] https://pastebin.canonical.com/111613/ [18:24] there's a way to select a node(machine) from maas, when deploying a service? [18:25] i'm planning to have different constraints in each node, so I would like to choose which one is going to be used [18:25] sebas5384: juju deploy mysql --to 3 [18:25] leftyfb: yeah, but thats when the machine is already mapped by the juju [18:26] you can also use constraints [18:26] https://juju.ubuntu.com/docs/reference-constraints.html [18:29] leftyfb: thanks!! [18:29] that was exactly what i was looking for [18:30] maas-name and maas-tags will be the solution [18:30] sebas5384: why do you want to select one specific node rather than just using e.g. cpu and memory constraints? [18:31] sarnold: because my client is going to provide the servers vms [18:31] the first battle was to implement juju and maybe maas [18:32] sarnold: IaaS is still going to be another battle for me [18:32] hehe [18:32] sebas5384: ahh, okay. :) I've just been happy to say "I want a big machine" or "I want a small machine" when deploying services in our cloud [18:32] juju > maas > node > X vmware vm [18:32] sebas5384: not knowing or caring how it happens is part of what I like :) [18:32] marcoceppi, how long is `juju clean local` supposed to take? [18:33] jcastro: 2-3 mins [18:33] ack [18:33] at most [18:33] jcastro: I've found if it's taking long, it's due to permissions on ~/juju [18:33] yeah sarnold! thats going to happen some day [18:33] er, ~/.juju [18:33] does it log anywhere? [18:33] but first i have to prove some concepts, it's a big and political company [18:33] jcastro: nope, hit CTRL+C and you'll see the errors [18:34] yep, correct! [18:34] marcoceppi: the script should use sudo or chown before trying to delete ~/.juju [18:34] juju clean ? [18:34] thats a new command? [18:34] leftyfb: it uses sudo when it needs to [18:34] is sudo juju clean a good idea? [18:34] jcastro: no [18:34] marcoceppi: when deleting ~/.juju, it needs to [18:35] I never got prompted to put in a sudo pw [18:35] leftyfb: bugs and pull requests welcome [18:35] but with the local provider .juju/local is owned by root [18:35] jcastro: it shouldn't be anymore [18:36] k, your example command is wrong too, I'll submit a PR for that tonight [18:37] i'm having bad luck with juju all around [18:38] my laptop won't bootstrap, and I just installed juju-local on the maas server in our lab here and it's stuck pending for juju-gui on machine 1 which says it's precise, even though the server is trusty. [18:39] also, according to jcastro's session the other day, using juju quickstart should put juju-gui onto machine 0 along with the bootstrap. That's not happening on the maas server [18:40] ooh!! https://github.com/juju/plugins/blob/master/juju-clean [18:41] sebas5384: you got an honorable mention in my UOS track [18:41] ooh yeah?? \o/ [18:41] leftyfb: have you specified a default-series in your ~/.juju/environments.yaml? [18:41] sebas5384: yeah - wrt the vagrant plugin we'll be working on [18:42] lazyPower: \o/ [18:42] great! [18:42] I feel like my track was meh, but in retrospect i'm feeling pretty beat up this week. Looking forward to recharging over the weekend. [18:42] get some Vitamin D, it wont be all rainy and blah outside [18:42] lazyPower: I didn't do so on the maas server ... to be honest I only installed juju on there to confirm that something was wrong with the setup on my laptop(confirmed) [18:42] lazyPower: i know that feeling ¬¬ [18:43] https://bugs.launchpad.net/juju-core/+bug/1329425 [18:43] <_mup_> Bug #1329425: Fatal out of memory error when bootstrapping on Azure [18:43] anyone see this before? [18:43] leftyfb: ok. That should correct the container series you see - but also during bootstrap its downloading a 200mb cloud image file - so it takes a couple of minutes to complete the first time. but i'm pretty sure you're aware of this since we checked /var/cache/lxc [18:43] leftyfb: where are you in terms of debug/nuking/retrying? [18:44] lazyPower: i've nuked several times ... give me a few to move out of the lab and go nuke again and start over [18:44] ack. ping when ready [18:44] sebas5384: so, now that its been phaux announced, we need to hammer out a time to do that sync we've been pending [18:45] yeah lazyPower absolutely! [18:46] lazyPower: my nuke involves much more hulk :) [18:46] leftyfb: oh man, sounds serious [18:47] lazyPower: what about 5pm EDT ? [18:47] friday [18:49] sebas5384: sure. I'm goign to send a calendar invite. [18:49] great!! [18:49] lazyPower: when is your track? [18:50] sebas5384: already wrapped about 10 minutes ago [18:50] bug 1329429 [18:50] <_mup_> Bug #1329429: local bootstrap fails

[18:50] oohhh :( [18:50] is it possible to bootstrap multiple environments at once? [18:51] lazyPower: invite acepted :) [18:52] jcastro: trying now to find out for you [18:52] jcastro: i have 2 running atm - one for maas another for amazon [18:53] juju switch maas && juju bootstrap -- in another byobu terminal - juju switch amazon && juju bootstrap [18:53] we'll see if things blow up, it'll take ~ 5 minutes to complete for maas [18:53] I might try a plugin, I want to do like "juju demo" and have it fire off a GUI in every environment in my yaml file or something [18:53] jcastro: that would be stupid awesome. do it [18:54] with some clever wrapping of quickstart you could even get it all fired up in your browser ready to rock n roll [18:55] yeah [18:55] lazyPower: this is prettmuch my hulking of juju: https://pastebin.canonical.com/111617/ [18:56] sinzui, have you ever seen anything like this perchance? https://bugs.launchpad.net/juju-core/+bug/1329425 [18:56] <_mup_> Bug #1329425: Fatal out of memory error when bootstrapping on Azure [18:57] leftyfb: even going so far as to remove teh packages. pretty thorough. [18:57] jcastro, no, Neither CI or I have ever experienced that [18:57] but that bug is old [18:57] lazyPower: obviously i'm missing something because it's not working from a "fresh" install [18:58] sinzui, old? I just filed it [18:58] You should have searched [18:58] it us the first bug in the azure-provider listing, bug 1250007 [18:58] <_mup_> Bug #1250007: Bootstrapping azure causes memory to fill

[19:00] jcastro, Did you use the cer instead of the pem? [19:01] This confusion also affects users of the python-sdk. MS is not very clear about what they expect [19:01] oh, I was using the .cer, fixing and trying with the pem [19:02] sinzui, that was it, I'll update the bug, it's bootstrapping fine, thanks! [19:03] sinzui, I will update the docs as frankban recommends in the master bug [19:03] jcastro, I learned that today, BTW. I just wrote a script to clean up the 10 terabytes of CI tests data juju left behind. I read a lot of support pages to learn that secret [19:04] hey lazyPower i watch your session! thanks for the mention! hehe [19:05] * lazyPower hides [19:06] was a great session lazyPower o/ [19:06] sorry that i missed hehe === CyberJacob|Away is now known as CyberJacob [19:09] I did however, do something productive - i fixed a bug in the docs - https://github.com/juju/docs/pull/120 [19:10] good one! that one was wrong since its there [19:10] xD [19:10] sinzui, mirrors on azure sick? http://pastebin.ubuntu.com/7635108/ [19:16] jcastro: oh! also - all is well with running bootstraps in paralell [19:19] is there a way to install juju server on a server of my choosing instead of a random server of juju's choosing? [19:22] d_`: is this OpenStack, AWS, HPCloud, what? [19:25] juju upgrade-charm of a local charm is giving me an error, saying that it can't find the charm [19:25] what can be? [19:25] sebas5384: did you specify --repository path on the juju upgrade-charm command? [19:26] hmmm [19:26] lazyPower: no ¬¬ [19:26] eg: juju upgrade-charm --repository=$HOME/charms myawesomecharm [19:26] sebas5384: i have a set of shell aliases when working with local repositories, because thats cumbersome to type [19:26] ahhh ok, i thought that was automatic [19:26] https://github.com/chuckbutler/dotfiles/blob/master/bashaliases <- sebas5384 [19:27] giving a look at lazyPower [19:27] oh hey look at that relic jrecycle in there... [19:27] * lazyPower updates [19:27] lazyPower, mine was a config issue, filed some bugs, all set now [19:27] other than getting an error with the mirrors [19:27] going to wait a few minutes and try again [19:27] ahhhh hokay [19:27] boooo on mirrors being wonky [19:27] jose: I'm using MAAS and it's just choosing a random physical server to install on [19:28] we have vmware here as well [19:28] lazyPower: nice, something like the juju clean plugin [19:28] he wants to specify where to deploy the bootstrap node it sounds like [19:28] i'm using Makefile for that hehe [19:29] could it be nice to have something like declaring repositories [19:29] or adding [19:30] for example in the environments.yaml file [19:30] repositories: [19:30] sebas5384: file a feature request [19:30] lazyPower: will do that :) === BradCrittenden is now known as bac === roadmr is now known as roadmr_afk [19:48] lazyPower: https://pastebin.canonical.com/111624/ [19:51] Declaring repositories in the environments.yaml -> https://github.com/juju/juju/issues/87 [19:54] leftyfb: thats really strange [19:57] lazyPower: you mean it's not broken by design? :) [19:58] leftyfb: i'm not sure what to mention here, as that issue is consistent with what the juju-clean script has thus far corrected. [19:59] but its progress now that its not complaining about simple-streams data [20:00] 2014-06-12 19:41:57 DEBUG juju.environs.simplestreams simplestreams.go:490 fetchData failed for "file:///home/leftyfb/.juju/local/storage/tools/streams/v1/mirrors.json": stat /home/leftyfb/.juju/local/storage/tools/streams/v1/mirrors.json: no such file or directory [20:00] 2014-06-12 19:41:57 DEBUG juju.environs.simplestreams simplestreams.go:567 no mirror index file found [20:00] 2014-06-12 19:41:57 DEBUG juju.environs.simplestreams simplestreams.go:548 no mirror information available for { }: mirror data for "com.ubuntu.juju:released:tools" not found [20:00] i stand corrected [20:00] i really dont know :( [20:01] leftyfb: have you pinged in #juju-dev to see if they know of any current issues similar to what you're seeing? [20:01] nope [20:01] That would be my next step, is to open a bug against this and ping with that in #juju-dev [20:14] i'm in an environment with limited internet access. when i deploy --to lxc:1, i see there is is a wget process trying to access cloud-images.ubuntu.com.. is there a right way to get around this problem in a closed environment? [20:15] here's the wget i see running http://pastebin.com/72q2BWsC [20:15] can i specify an internal location for the images perhaps? === roadmr_afk is now known as roadmr === vladk is now known as vladk|offline [21:00] hazmat: have you tried the docean plugin on Mac OSX? [21:06] lazyPower: fixed it [21:26] do we have a collective noun for "juju environments" ? [21:26] * thumper was thinking "a collective of environments" [21:26] but perhaps that is too borg [21:26] thumper: environments? [21:27] marcoceppi: that isn't a collective noun [21:27] think "a murder of crows" [21:27] "a horde of hamsters" [21:27] a bushel of environments [21:27] heh [21:27] a gaggle of environments [21:28] * thumper briefly thought "group" [21:28] but group has other meanings [21:29] a plethora of environments [21:29] a plague of environments... [21:29] thumper: cull? [21:29] "a host of environments"? [21:29] hmm... [21:30] I'm trying to avoid overloaded terms... [21:30] so trying not to have a collective noun that relates to other well known terms [21:30] you could call it a charm of environments [21:30] so I guess avoiding: host, group, insert-other-techy-term [21:30] that overloads the term "charm" [21:30] which I'm trying to avoid [21:30] I know <3 [21:30] :) [21:31] leftyfb: nice! what was teh fix? [21:31] what about "an illusion of environments" [21:31] apparently it is "an illusion of magicians" [21:31] https://bugs.launchpad.net/juju-core/+bug/1329480 [21:31] <_mup_> Bug #1329480: Cannot create juju-run symlink if /usr/local/bin doesn't exist [21:31] thumper: howa bout "a lot of environments" [21:31] or, better yet, alot [21:31] just make up our own [21:32] leftyfb: interesting... i would have assumed /usr/local/bin would just exist. [21:32] devs should never assume :) [21:32] marcoceppi: how about "an ensemble of environments" [21:32] A Quafffel of Environments [21:32] leftyfb: I assumed /usr/local/bin existed [21:32] * thumper thinks ensemble is cute [21:32] thumper: i see whatchya did there [21:33] and a nod to pre-juju [21:33] done [21:33] ensemble of environments it is [21:33] add it to the docs! [21:33] it's official [21:37] is it possible using juju-gui to install to a specific container? [21:38] Is there a way to ask juju for the public IP address instead of the public-address name? [21:39] cory_fu: which provider? [21:40] I mean a general way for a charm to get the public IP as opposed to the public address [21:41] cory_fu: depending on the substrate, some respodn with just teh ip. other's i've had to run through dig +short to get the proper IP [21:41] Ok, that's what the charm is doing currently; I was just hoping there was a better way [21:41] there's an open bug on this somewhere - i'm pretty sure matt williams opened one during our sync on the DNS charm [21:42] it seems when you deploy services and then destroy them, juju-gui will leave those containers running and spin up new containers when deploying new services [21:42] leftyfb: thats correct. by design juju will not destroy the machine after the service has been destroyed. [21:43] lazyPower: and no way to re-utilize that machine for future deployments using juju-gui? [21:43] leftyfb: not using the gui, no [21:43] bug worthy? [21:44] leftyfb: there's an effort to bring machine-view to the juju-gui, which will alleviate some of this in the future, along with easing deployment of KVM/LXC units to a server for density [21:44] cool [21:45] cory_fu: i'm unable to locate the bug if one was opened. so it may not have been raised. [21:46] No biggie. I don't think it actually mattered if I had an IP in this current case, actually, and, while it does in the other case and it would be nice to clean it up, it works currently, so... [21:46] cory_fu: wait i found it - https://bugs.launchpad.net/juju-core/+bug/1308374 [21:46] <_mup_> Bug #1308374: unit-get public-address on EC2 returns split horizon DNS

[21:47] lazyPower: how about specifying distro version (precise/trusty) for deployments? Everything seems to be precise even though I have default-series: trusty in my ~/.juju/environments.yaml [21:47] leftyfb: our trusty charmstore isn't very big [21:47] http://manage.jujucharms.com/charms/trusty [21:49] leftyfb: we have a rather large queue to get our existing prceise charms tested, updated to meet current charm store best practices, and unit/integration tested for inclusion into trusty - if you'd like to volunteer to help we could certainly use a helping hand - current work items are being managed in this spreadsheet - https://docs.google.com/a/canonical.com/spreadsheet/ccc?key=0Aia4W3c4fbL-dGs4SVBJMGRIdnlSMWhzSmo3WE1mZ1E&usp=drive_web#gid=0 [21:50] lazyPower: once I nail down juju, I might just offer some assistance in that effort [21:51] I can't even remove a service at this point [21:51] leftyfb: juju destroy-service [21:51] juju destroy-machine # [21:51] lazyPower: yeah, it'd be nice if that worked [21:51] is the charm in an error state? [21:51] mysql: [21:51] charm: cs:precise/mysql-45 [21:51] exposed: false [21:51] life: dying [21:51] units: [21:51] mysql/0: [21:51] agent-state: error [21:51] agent-state-info: 'hook failed: "start"' [21:52] agent-version: 1.18.4.1 [21:52] life: dying [21:52] machine: "4" [21:52] public-address: 10.0.2.31 [21:52] it failed to start [21:52] clicking on "resolve" or "retry" did nothing [21:52] you'll need to resolve the hook error before it will allow any of the other events to run. its goign to hang in that error state until you resolve it [21:52] with no output on the juju-gui as to what the issue was [21:52] how do I find out what's wrong? [21:53] you can inspect juju debug-log -n 500 to find the issue if the logs caught it - otherwise you'll want to investigate in running juju debug-hooks to run the hook interactively to see why it failed [21:53] leftyfb: if debug-log doesn't tell you anything - https://juju.ubuntu.com/docs/authors-hook-debug.html [21:55] leftyfb: also - note - since its in state dying, even if you resolve it successfully - your service is queued for destruction, and will promptly be destroyed once the issue has been resolved. [21:55] lazyPower: you know a customer is jumping through hoops at this point right? All I did was drag mysql to my canvas and clicked deply. Now the customer is looking through log files and searching for what a "hook-name" means [21:56] leftyfb: unfortunately - I dont have a better answer at this time. [21:56] what is the unit name? [21:56] I tried both mysql and cs:precise/mysql-45 [21:57] mysql/0 - which is shown directly under units [21:57] in juju-status [21:57] for each node in a deployment, get increment by 1, so your second mysql unit if the service is scaled would be mysql/1 [21:58] leftyfb: if you're deploying MySQL on the local provider - there is a known bug with the charm that will prevent it from deploying successfully [21:58] :/ [21:59] it's not on the bootstrap machine if that's what you mean by local provider [21:59] if you mean a separate lxc container, then that's how it's setup at the moment .... what's the resolution? [22:02] leftyfb: LXC is a best attempt at emulating clouds on your local machine - so there are some subtle differences between it and a true cloud provider that sometimes cause unintended side effects with teh charm. I'm looking for the solution now, its been a while since i've used MySQL on the local provider. [22:02] i'm seeing possible memory exhaustion, that ring a bell? [22:04] juju set mysql dataset-size='512M' [22:04] juju resolved -r mysql/0 [22:05] that did it [22:06] i'll mess with this more tomorrow/later [22:06]