[00:13] Question: I just installed juju on 16.04 from repo (juju 2.0 beta15) and when I try to connect to the gui, it says connecting to Juju model and never goes past. Nothing in the logs looks suspicious other than in the machine-0.log there is an entry about reconfiguring lxd using dpkg and setup lxcbr0 which I do, but still nothing. Rebooted the controller and I had to tell Firefox to delete the SSL certificate for the server and [00:13] have it re-issue it so that I could connect again, but same results. [00:13] wililupy: add juju/stable first ... [00:13] bdx even on xenial? [00:13] wililupy: `sudo add-apt-repository ppa:juju/stable` [00:14] wililupy: yea ... not for much longer I think .. [00:22] bdx: I'm updating it right now. Do I have to re-bootstrap? [00:22] wililupy: yes [00:22] wililupy: `rm -rf ~/.local/share/juju` too [00:22] bdx: ack. [00:47] bdx: That fixed it. Thank you so much. I eventually probably would have done that, but I was hoping not having to add a ppa. === zz_CyberJacob is now known as CyberJacob [08:55] magicaltrout: yep [08:56] unlucky === ant_ is now known as Guest44024 [17:22] lutostag, beisner, marcoceppi: where might one find the interfaces for the charm-jenkins charm? [17:25] marcoceppi, bdx, it looks like they're not in http://interfaces.juju.solutions :-/ [17:27] marcoceppi, bdx, i previously built using https://github.com/freeekanayaka/interface-jenkins-slave, which really should make it into jenkins upstream along side the jenkins-charm layer repo. [17:29] beisner: nice, thanks [17:29] looks like free has the other interfaces stashed there too :-) [17:34] cory_fu, so the tweaks to bundletester config did not explode in the gate - https://review.openstack.org/#/c/395574/ [17:35] jamespage: That's good. We didn't get that rabbitmq branch into CWR, though, unfortunately. Is that still something you need? [17:36] cory_fu, it would be helpful to get this right first time - however if its alot of effort, we can probably get away with landing the changes and testing in stable, but that's not ideal [17:37] cory_fu: jamespage: the rabbit charm made it into cwr, just not in time for the overnight runs. i'm pretty sure it'll be picked up when they kick off today. [17:38] kwmonroe, based on my test charm under ~james-page right? [17:38] correct jamespage [17:38] cs:~james-page/rabbitmq-server-requirement [17:38] Yeah, sorry, I meant what kwmonroe said [17:39] jamespage: It will be run automatically this evening, but if need be, we could kick it off manually sooner, though I think it would still take a while before it actually runs [17:39] That CI infra is pretty over-committed, from what I can tell [17:40] cory_fu, I just synced it up so I'll look am tomorrow [17:40] tvansteenburgh1: it appears i'm hosting a timtimtim model in azure-east. pretty sure that was for long-ago k8s test. cool if i destroy that? [17:40] Ok [17:40] kwmonroe: yes [17:40] thx === tvansteenburgh1 is now known as tvansteenburgh [17:41] woot thanks jamespage cory_fu [18:11] Does anyone have a sample of yaml config file to be used during bootstrap. Cannot find any reference on the docs, only for charms. [18:25] deanman - its just a key: value map. http://paste.ubuntu.com/23457321/ [18:26] lazyPower: ok got it! Thanks [18:47] lazyPower: Can you put comments and/or environmental variables as values? [18:47] in yaml? [18:48] no, you would need to render it through a template wizard like envtemplate or something. [18:48] however, comments should be just fine, you can prefix a comment in yaml with # [18:48] lazyPower: ok thats seems reasonable! [20:11] Is it possible to `juju run` using a command that takes arguments? A simple `tail -f` seems to give an error with "flag provided but not defined: -f" [20:13] deanman - you'll need to quote the command [20:13] eg: juju run --application foobar "ls -al && whoami" [20:16] any way to move bootstrap to a different ip/machine on 1.25.x? Our bootstrap node has run out of disk space. (Or a way to purge unused data from /var/lib/juju would work too)? [20:17] lazyPower: Shouldn't this work ? juju run -m controller --machine 0 "tail -f /var/log/juju/machine-0.log" [20:17] lutostag - thats rough, i think the option is to either go HA (probably not going to solve disk space issues) or to purge the cache/logs/whatever-is-using-the-diskspace [20:17] it just hangs [20:17] deanman - its async i think,s o you wont see the output until the command terminates [20:17] deanman - you should probably juju ssh into the unit if you need to tail logs. [20:18] ok lazyPower [20:19] lazyPower: is /var/lib/juju/charm-get-cache safe to boot into the bit-bucket in the sky? [20:20] i think so, i'm fairly certain it will just replace whatever it needs that it doesn't find in the cache [20:20] *but* [20:20] with that being said, i haven't done it myself, so it might break things in unexpected ways [20:20] lutostag - i would highly recommend we file a bug about this so we can add an operational action to the controller to purge temporary/unneccesary data in the event this happens [20:20] lazyPower: ack will find/create a bug [20:23] lazyPower: somehow even juju run -m controller --machine 0 "ls -al" hangs there. [20:23] mattrae ping [20:36] deanman: does `juju ssh -m controller 0` work? [20:38] deanman - yeah that doesn't sound good. give babbageclunk's suggestion a go [20:48] babbageclunk: it does but i have since restarted the VM and now even run works [20:48] deanman: sounds like sshd was having a bad time. [20:49] babbageclunk: it does take though like 5 seconds for a whoami [20:49] babbageclunk: is that normal? [20:52] deanman: no that sounds very slow... [20:54] babbageclunk: http://paste.ubuntu.com/23457920/ [20:54] seems like the previous command "tail -f" makes it unresponsive [20:56] deanman: yeah, I see the same behaviour - can you file a bug for that? [20:57] `juju ssh` makes it recover somehow [20:57] sure babbageclunk ! [20:58] deanman: Thanks! [20:59] babbageclunk: can you `time juju run -m controller --machine - "whoami" for me and tell me what's your average? [21:00] i get 2-5 secs [21:01] deanman: yeah, I get the same - 5 seconds sounded slow until I tried it. [21:08] babbageclunk: https://bugs.launchpad.net/juju/+bug/1640938 [21:08] Bug #1640938: `juju run` becomes unresponsive in consecutive runs when previous command passed is `tail -f` [21:09] deanman: great thanks [21:09] Cheers! [21:15] babbageclunk - that tail -f doesn't exit, and i'm pretty sure that its just hanging the ssh session until timeout [21:15] deanman - humor me and try the tail with something like a -n 100 and omit the -f [21:16] lazyPower: Oh, I think that's right, I'm just surprised that killing the run command doesn't kill the tail. [21:25] Having trouble with image-metadata bootstrapping private cloud, have meta data json in ~/simplestreams as described in "getting" started but no matter how I feed it in keep getting [21:25] ERROR failed to bootstrap model: no image metadata found [21:25] tried both --metadata-source=~/simplestreams and webserver using --config image-metadata-url=https://people.csail.mit.edu/jon/simplestreams amoung other things....Help? [21:25] oh juju-2.0 on xenial [21:30] jproulx - i'm not as familiar with the clouds that require manual provisioning of simplestreams, however if you dont get an answer here in a timely fashion, the mailing list is an excellent resource for help with problems like that. https://lists.ubuntu.com/mailman/listinfo/juju [21:31] lazyPower: thanks for the pointer [21:40] lazyPower: hey saw your ping :) [21:40] mattrae - found the answer already, thanks for pong'ing [21:40] cool no problem :) === redir is now known as redir_excorcise === redir_excorcise is now known as redir === CyberJacob is now known as zz_CyberJacob