[00:02] bdx: hey, you found a cloud init bug, do you have the link? [00:03] marcoceppi: https://bugs.launchpad.net/ubuntu/+source/snapd/+bug/1621229 [00:03] Bug #1621229: snap upgrade to 2.14.2~16.04 in xenial lxc hangs [00:03] lutostag: this is actually a huge problem, do you know if it has an eta? === thumper is now known as thumper-lunch [00:04] marcoceppi: no, its caused by the sru for snap, should be fixed with next daily though [00:04] lutostag: so, will the daily be what people normally get? [00:05] every juju deploy in a lxd container of xenial is broken right now, and we're 4 days away from a charmer summit [00:05] thats what I get with default juju commands -- and how I ran into it [00:05] bdx [00:05] 's workaround FTM: http://paste.ubuntu.com/23147632/ [00:06] lutostag bdx that will only hold true until xenial refreshes it's image cache [00:06] but yes, workarounds for something supposed to be super stable is not a nice place to be so close to so many demos [00:06] lxd's or some other cache? [00:08] lxd's, every ~2 weeks or whatever, when our cloud team releases new images, LXD's image cache gets updated on all those who have network access and enabled it (enabled by default if you did `lxc launch ubuntu:16.04` [00:12] marcoceppi: ah, I usually pull from https://images.linuxcontainers.org not https://cloud-images.ubuntu.com/releases, so yeah more of an impact than I thought [00:13] Sept 1 was the last image update I got for Xenial [00:14] so yes, iiuc it will hit us as long as we have snap as an updating package w/ cloud-init (until update for xenial -- and I'm pulling from release NOT daily) which makes it a long-lasting one [00:14] Aug 30, 2016 at 12:00am was when it was updated simplestreams server-side [00:15] I believe smoser had a workaround in-mind cloud-init side before he EOD'd [00:17] looks like it got escalated correctly to mgmt -- but jgrimm is who you throw hell-fire at I believe to ensure timely fix [00:22] no need for hell fire, a bug like this with no "severity" concerns me [00:24] its parent has critical as of ~15mins ago [00:26] lutostag: yeah, I mean that was a recent change :) [00:44] o/ thumper-lunch [00:44] hey, this might seem like a dumb question, but why would the color in `juju status` not work when doing `watch juju status`? [00:48] jcastro: because the process detects if stdout has redirected or not and doesn't output color if it has. same as `watch ls -l` [00:48] jcastro: use watch -c and... hopefully juju status has a force color option? [00:49] jcastro: yeah, watch -c juju status --color=true should work [00:49] yep, got it, thanks [00:50] jcastro: once a decade i actually help you. ;] [00:50] jcastro: and it is always rather trivial. [00:52] yeah it's just that colors really help when doing workshops, so like, we should totally use that in Pasadena [00:52] jrwren: are you going to pasadena? [00:53] watch? [00:53] you all need to do this :) [00:54] http://paste.ubuntu.com/23148247/ [00:55] jcastro: ^ ;) [00:56] jcastro: yes, I am going. === thumper-lunch is now known as thumper [01:24] jcastro: hey [01:24] jcastro: do this `watch -c juju status --color` [01:26] actually I see that jrwren has already sorted you out [05:18] magicaltrout, bdx : http://interface-pgsql.readthedocs.io/en/stable/requires.html has some examples. I needed to break backwards compatibility with the example pgsql interface when I took it over, as it didn't support a lot of my real world deploys. [05:19] hatch: Thank you very much. We will try this. === frankban|afk is now known as frankban [07:38] I am stuck with the following problem [07:38] http://askubuntu.com/questions/821804/openstack-with-landscape-install-fails [07:53] stokachu: could u take a look at my question above? :) [08:51] boom [08:51] working apache drill with mysql and postgresql [08:51] excellent [11:05] https://ibin.co/2uMOTf29nfWf.png [11:05] https://ibin.co/2uMOYLOeKkkP.png [11:05] https://ibin.co/2uMOc5gdxZzh.png [11:05] this is pretty cool [11:05] juju charms hook up to drill [11:06] then let me run a query across multiple charm based data sources without hacking around === dpm_ is now known as dpm [12:02] magicaltrout: that's awesome [12:03] thanks babbageclunk [12:04] part of my requirements to finally get some decent analytics tools other than Zeppelin running over the top the various data sources juju provides [12:04] but also allow reasonably easy data federation across a bunch of different sources for easier ETL etc [12:32] magicaltrout: sweee6 [12:37] aye marcoceppi coming along [12:37] I've now gotta fix up my saiku charm and create a drill interface so they can talk and we'll be good [12:40] http://spicule.co.uk/2016/09/08/juju-drill-amazingness.html [13:14] Hi. We have developed a "cinder-storagedriver" charm with one of our storage array. Our charm will be similar to "cinder-ceph". Already our charm is in public charm store. We want to integrate our charm to "OpenStack Base bundle" and then we want to push the bundle to Charm store. So How can I integrate my charm to OpenStack Base bundle? [13:14] To push the bundle to charm store, Can I follow the same procedure as we followed for pushing our charm to store?. [13:15] And Once we have the Bundle in the public Charm store , before deploying the bundle can we set configuration parameters externally using "juju deploy bundle-name set-config key=value,......" command? Because while integrating our charm with "OpenStack base bundle" we will provide default configuration values[Like San IP, San Password ,San user,....] to our charm. [13:15] Please anyone provide me some solution for this. [13:15] rock: you'll need to deploy the openstack base bundle, add your stuff, and push that bundle back to the charm store as your own bundle [13:17] in the mean time, if you actually want your cham included in the open stack base bundle, you'll need to get some tests written for it and it into the review queue, reviewed by a charmer and promulgated I suspect [13:17] i hit this bug just now, not sure how to get around it though, tips? https://bugs.launchpad.net/ubuntu/+source/snapd/+bug/1621336 [13:17] Bug #1621336: snapd.boot-ok.service hangs eternally on cloud image upgrades [13:20] magicaltrout: Hi. How can I integrate my charm in OpenStack base bundle?. [13:20] I just told you [13:24] magicaltrout: didn't get fully. I have one option. Tell me that is correct (or) not. I will download Openstack bundle and add our charm details in bundle.yaml file. Then I will push this bundle to store as our own bundle. And then I will try to deploy that bundle by taking from store. [13:24] correct [13:25] magicaltrout: Then i have a question. Once we have the Bundle in the public Charm store , before deploying the bundle can we set configuration parameters externally using "juju deploy bundle-name set-config key=value,......" command? Because while integrating our charm with "OpenStack base bundle" we will provide default configuration values[Like San IP, San Password ,San user,....] to our charm. [13:27] dunno [13:27] you can certainly do that in the GUI, I assume you can do it on the command line [13:27] never tried it though [13:28] rock: no, there's not currently a way to override config in the cli [13:28] okay [13:28] i lied [13:28] rock: magicaltrout it doens't have the idea of "uncomitted" that the GUI has at this time. It's something we're interested in looking into, but the cli is more direct and immediate without a "deploy" go button like the GUI [13:29] yeah i did wonder about the staging aspectr [13:29] -r [13:32] rick_h_: we did this for changing the default config for our charm.1) $sudo juju deploy cs:~siva9296/kaminario-openstack-5 san1 2)$sudo juju set-config san1 san-ip=192.168.0.1 san-user=admin1 san-password=password1 storage-protocol=iscsi os-version=liberty 3)$sudo juju add-relation san1 cinder [13:33] rick_h_: It worked for our charm. [13:33] rock: why do you need to use sudo? [13:33] rock: agree that you can set the config after the fact ok [13:34] rick_h_: Actually we no need. While doing juju bundle setup we run all commands with sudo. [13:35] what rick_h_ you don't grant root perms to all your workforce? ;) [13:38] rick_h_: While integrating our charm to openstck base bundle , we will give default config values to my charm. When user want to deploy our bundle then he will set actual config parameters using set-config . This what we want to do. [13:39] so you can do that in the GUI rock, as explained, you can't do it on the commandline at bundle level [13:39] you'd have to deploy the bundle then set the config for the charm you need to configure after the bundle starts spinning up [13:39] rock: one thing you might want to do, is have the charm ship with empty values. Then in the charm code say "If I have no configuration for SAN, block the charm and warn the user I need that information before proceeding" [13:40] rock: that way the user will get the message after deployment, and the deploy won't break because of bad configuration === ant__ is now known as antdillon [13:50] for the firs time in a while I'm actually excited about some Saiku stuff we're finishing off. The Juju connectivity and schema generation will make Analysis over various data sources so much more effective, and take next to no time to setup thanks to Juju [13:52] magicaltrout: sounds like you just signed up for a lightning talk ;) [13:52] funny that, there appears to be a BI related one with my name next to it... [13:53] oic [13:54] these are the missing bits for the lightning talk but also a driver to get me to finally get the next release over the line because it will make demo's and building these data platforms much easier [13:55] so I figured having an enforced deadline helps focus the mind ;) [13:55] but also from a juju point of view, having a user friendly tool to run over HDFS, MySQL, whatever is nice as well [13:56] I can't expect users to use Zeppelin, as nice as it is they don't know how to write queries [13:56] marcoceppi: Ok. thank you. Finally, I got the point like- If I integrate our charm to bundle and push that bundle to store as our own bundle, Then while deploying bundle using "juju deploy bundle-name" we can't set configration values right. Once after having the bundle deployment only we can set config parameter values right. [13:57] rock: exactly, after you type `juju deploy bundle-name` anyone can set configuration either via the GUi or CLI [14:06] marcoceppi: Thank you. I have one more question. [14:07] marcoceppi: This question related to basic charm creation. pasted issue info. http://paste.openstack.org/show/569211/ [14:12] marcoceppi: For juju version 2.0-beta15-xenial-amd64, If I run the command "$charm create test" Then it was creating basic charm template with "hooks folder". and "$juju charm create test' this command is unrecognized. followed https://jujucharms.com/docs/2.0/authors-charm-writing [14:13] marcoceppi: sorry. $charm create test" Then it was not creating basic charm template with "hooks folder". [14:13] rock: charm command and juju command are two separate tools [14:14] rock: odds are, in trusty, you're installing a very old version of the tool. Charm creation has changed a lot since 2014. One change is that `juju charm` is no longer a command, and it's just the `charm` command. The second change is a move to layered charms. This is where you write less code and use `charm build` to compile a charm (create the hooks directory, etc) [14:15] rock: this is the updated documentation for this: https://jujucharms.com/docs/stable/developer-getting-started [14:15] rock: you can install the latest charm command on trusty using this: https://jujucharms.com/docs/stable/tools-charm-tools [14:16] marcoceppi: OK. Thank you. [14:18] rock: very excited to see your charm coming together! Let us know if you have any other questions [14:30] marcoceppi: Thank you very much for your support. Right now I have cleared my questions. We will move forward now. If we stuck at any place. We will ask you. Thank you. [14:34] Anita [14:35] Hi Matt [16:24] aisrael: lazyPower: here's cory_fu's jdo snippet: http://paste.ubuntu.com/23150665/ [16:24] Very nice, thanks! [16:26] gracias! [16:37] kjackal: https://github.com/juju-solutions/interface-spark [16:41] kwmonroe cory_fu why not something like this? http://paste.ubuntu.com/23150892/ [16:41] seamless integration ;) [16:44] nice marcoceppi [16:45] and a pager for juju status [16:47] sweet snippet [16:48] marcoceppi: I like the status helper. Gonna have to add that. [16:49] marcoceppi: What version of juju supports colored status? I get a "flag not defined" error from beta16 [16:52] http://askubuntu.com/questions/821804/openstack-with-landscape-install-fails [16:52] does someone has an idea? [16:53] cory_fu: think it's 17 === frankban is now known as frankban|afk [16:58] PCdude - that seems odd, but i'll relay this AU question to the landscape dudes. [16:58] s/dudes/team/ [17:05] I tried to ping them yesterday [17:05] dpb1: any of you folks around? ^^^ [17:26] jcastro - looks more like a conjure issue when i look deeper at the Au question, i've pinged stokachu re ^ [17:26] ack [17:28] lazyPower, aisrael: This the thing? https://www.amazon.com/Reusable-Rubber-eLander-Assorted-Colors/dp/B01ID1XXE2/ref=sr_1_9?s=electronics&ie=UTF8&qid=1473355654&sr=1-9&keywords=twists [17:28] yep, same company/ties [17:28] mine are the shorter ones, came in an 8 pack [17:29] lazyPower: I was disconnected for a moment, so I dont know if u wrote a reply :) [17:29] PCdude not as of yet, the primary author of that utility is afk. but we've pinged for some help for you [17:30] lazyPower: great! I will wait in this channel for the next couple of hours [17:30] lazyPower: not always behind PC, but online [17:30] PCdude - that works, but i would monitor the AU question, as thats what we sent over. [17:31] lazyPower: ah ok, that a good idea. I will keep an eye on that one too [18:28] PCdude, ive posted a response, if you need more assistance beyond what the community can provide via askubuntu I urge you to look into Ubuntu Advantage [18:31] stokachu: thanks for ur reply. I tried that already with success. So the bootstrap and the the deployment succeed. I dont know if u have followed the discussion in the maas channel, but the problem is getting more clear [18:31] PCdude, no i haven't [18:31] what turns out that the problem is related to spinning up machines with LXC by JUJU. Apparently there goes something wrong [18:32] stokachu: I will give a short overview [18:32] http://pastebin.com/raw/Lh1uTDQt [18:32] that is the output of "sudo lxc-ls --fancy" [18:33] ok that doesn't really tell me anything [18:33] stokachu: http://pastebin.com/raw/AZY2ZHiV [18:34] there is output of "JUJU_HOME=~/.cloud-install/juju juju status" [18:34] JUJU can for some reason not spin up machines at it should be [18:35] on advice of other the creator of the "openstack-installer" for ubuntu, I should head over to the JUJU channel for more help [18:35] PCdude: there's a known issue with bringing up instances today. There's a bug in a package update that causes instances to not get through apt-get update/upgrade properly when brought up. [18:35] PCdude: that's being corrected, in the mean time you have to set the auto upgrade to false on juju models for now. Looks like you're on juju 1.25 [18:37] rick_h_: yes indeed, I am using JUJU 1.25. Where can I set that option? which file? this would be great if it works :) [18:37] PCdude: looking, I did it for 2.0 today so trying to remember how to set it in 1.25 [18:38] PCdude: add enable-os-refresh-update: false and enable-os-upgrade: false to environments.yaml manually [18:38] PCdude: check out https://lists.ubuntu.com/archives/juju/2016-September/007845.html for hte background and then add those keys to your juju config in environments.yaml [18:39] what jrwren said [18:39] what rick_h_ said [18:39] PCdude: and if that works to help those lxc containers come up the problem should be working itself out here shortly [18:40] rick_h_: just to be sure, I am not using xenial, but I am using trusty. That makes this solution still valid? just trying to be sure [18:43] PCdude: oh, hmmm...good point, that's probably not true then. [18:43] PCdude: have to check the lxc logs then and see what's up on them [18:44] rick_h_: I never used LXC so which log u wanna see (command would be handy too haha) [18:50] PCdude: can you ssh/connect to the lxc containers and get /var/log/juju/* logs from there? [18:52] rick_h_: http://askubuntu.com/questions/821804/openstack-with-landscape-install-fails [18:52] that is mine question in total [18:52] the second block is the output of what u are asking on the node that runs the LXC containers [18:52] is that what u are looking for? [18:53] *my [18:54] ugh wait [18:54] are you running out of dhcp addresses? [18:54] what is your dhcp and static ranges defined in maas [18:54] PCdude, ^ [18:57] stokachu: http://i.imgur.com/rg0JcaO.jpg [18:57] 50 for dynamic and 50 for static [18:57] I seriously hope that is enough :) [18:58] PCdude, well if you've run the installation several times in a short period [18:59] stokachu: good point, I have everything running in ESXI and heavily use the snapshot function. So, every install seems to be the first for the system [19:01] PCdude, no idea didnt realize you were running this all in vmware as that info was left out in the askubuntu question [19:02] PCdude, so you said you could juju bootstrap outside of running the installer? [19:02] did you deploy anything? [19:02] and did you deploy anything to a container within a machine? [19:04] stokachu: yes, I did the bootstrap and it succeeded also I deployed ubuntu with it. also that succeeded, as I could login and use the machine. I have not issued any other commands [19:04] stokachu: no I did not deployed anything to a container within the machine [19:04] PCdude, you did a juju deploy ubuntu? [19:05] stokachu: correct [19:05] PCdude, do that again and then run juju deploy wordpress --to lxc:1 [19:05] juju deploy mysql --to lxc:1 [19:06] stokachu: ok, np. I have to set some things back and bootstrap again. So this can take some minutes [19:06] I will report back when I am done [19:06] k [19:06] dont tear it down if it fails [19:09] stokachu: I will try :) its bootstrapping rn [19:25] have you guys heard about, or know about this -> https://postimg.org/image/qiaid5uo7/ [19:25] its been blocking me from launching any xenial ami for weeks now [19:26] stokachu: I have to issue that command on the controller not the node right? [19:26] thought I would try again today, and it still exists [19:26] stokachu: the bootstrap and the deployment are both done and succeeded [19:27] bdx: no, there's a know issue from last night to today, but nothing going on weeks we're aware of. [19:27] bdx: have you reached out to gaughen and company on this? [19:28] no, I just thought I would bring it up now [19:28] * gaughen looks up [19:28] bdx: We haven't been seeing widespread issues launching 16.04 AMIs. [19:28] bdx: gotcha, yea not aware ofthat little badge and looks like we've got gaughen's attention so hopefully she's got a better idea there. [19:29] PCdude, what commands did you run [19:29] rick_h_: great, thx [19:30] "juju bootstrap" and "juju deploy ubuntu -n 1" [19:30] stokachu: thats correct right? [19:30] PCdude, now try juju deploy mysql --to lxc:1 [19:30] bdx, do you have any more specific info? [19:31] stokachu: ok, I did that and it succeeded I think (at least it gave me no error and a new line in the CLI) [19:31] gaughen: I have a report of the same thing happening with trusty images from a coworker all last week [19:31] PCdude, so you have a container with an IP? [19:32] gaughen: trusty amis* [19:32] bdx: Ah, I didn't immediately recognise that as the OpsWorks dashboard. [19:32] bdx: There _are_ known issues there at the moment, but I don't believe that it's been going on for weeks. [19:33] bdx, there is a known issue (as of yesterday) for opworks where instances are hanging during boot [19:33] stokachu: uhm, I dont use LXD, what command should I use? [19:33] but it does not impact Trusty, and hasn't been an issue for weeks [19:33] PCdude, huh? [19:33] PCdude, the command is `juju deploy mysql --to lxc:1` [19:33] stokachu: to see containers with IP address, so I can reach them? [19:34] stokachu: yes I did that, but after it to check it worked [19:34] you can use the lxc-list --fancy [19:34] or juju status [19:34] Odd_Bloke: it doesn't matter if its launched in opsworks ... between another dev and myself we probably launched 100 amis over 4 aws accounts in different regions, using different amis and they would immediately terminate on boot [19:35] bdx, would you file a bug here https://bugs.launchpad.net/cloud-images with as much detail as possible [19:35] bdx: There was an outage (unrelated to OpsWorks) on Friday for instance-store AMIs. [19:35] but when in the opsworks console, you get that message [19:35] including steps for me to repro [19:35] bdx: But that was only an issue for a few hours before we recovered. [19:36] stokachu: juju status: http://pastebin.com/raw/XdvKet93 [19:36] bdx: So, yeah, as gaughen says, a bug with reproduction instructions (including AMI IDs) would be much appreciated. :) [19:36] ahhh, now that I'm bitching, I just got a xenial ami to launch [19:36] totally [19:37] PCdude, and what does sudo lxc-list --fancy show [19:37] not through opsworks though [19:37] by manual provisioning [19:38] Yeah, the OpsWorks issue is ongoing; the root cause is https://bugs.launchpad.net/cloud-init/+bug/1576692 [19:38] Bug #1576692: fully support package installation in systemd

[19:38] I'll fire out a bug for this by eod [19:38] ok, good to know [19:38] stokachu: WTF, command not find, let me triple check this. this cant be right! [19:38] PCdude, sorry lxc-ls --fancy [19:39] stokachu: ah ok, that has a result haha [19:39] stokachu: http://pastebin.com/raw/bxAaxwFN [19:40] stokachu: that seems ok right? [19:40] PCdude, no it's not getting an IP [19:41] stokachu: ah ok ,yes that should be there. At least we know for sure now, that its not landscape or anything else but just JUJU. [19:42] PCdude, well juju just calls out to lxc to create the container [19:42] this could be your MAAS server, try expanding the static IP range pool [19:43] but yes it isn't landscape, or the openstack-installer, and im pretty sure it isn't juju [19:43] stokachu: ah ok, well how big should I make it go from 50 to 100? [19:43] yea try 100 [19:44] PCdude, once you do that in maas, try to deploy a new charm like, juju deploy wordpress --to lxc:1 [19:45] stokachu: ok, it now has 100 IP addresses lets try again [19:46] stokachu: still no IP address... [19:47] PCdude, juju ssh 1; sudo pastebinit /var/log/juju/unit-wordpress-0.log [19:47] you can also lxc-attach to one of those containers [19:48] and look at /var/log/cloud-init [19:48] stokachu: wordpress-0.log does not exsists. I only got machine-1.log and unit-ubuntu-0.log [19:48] PCdude, pastebinit machine-1.log [19:49] PCdude, also try lxc-attach -n juju-machine-1-lxc-0 [19:49] and paste /var/log/cloud-init* [19:49] http://paste.ubuntu.com/23151482/ [19:49] ok that is /var/log/cloud-init.log [19:50] http://paste.ubuntu.com/23151485/ [19:50] that is /var/log/juju/machine-1.log [19:51] I did lxc-attach -n juju-machine-1-lxc-0 [19:52] it indeed has no IP [19:52] not sure what 2016-09-08 19:24:37 INFO juju.networker networker.go:163 networker is disabled - not starting on machine "machine-1" means [19:52] rick_h_, ^ [19:53] jose: is your couchbase charm online somewhere i can look at it? [19:53] stokachu: it can ping 127.0.0.1 but thats it. thats of course the loop back, but still nothing is crippled to start with [19:53] * mhall119 wonders if jcastro would mind if he converted jose from a charmer to a snapper at UbuConEU [19:54] stokachu: not sure [19:54] kjackal_, petevg, kwmonroe: https://github.com/juju-solutions/interface-zeppelin [19:54] Please review [19:57] PCdude, is the 192.168.1.x a private network? [20:00] * mhall119 is pretty sure that's a private network address block [20:00] mhall119: launchpad but i'm pushing another iteration soon [20:01] mhall119, thanks for taking over support [20:01] stokachu: darnit :( [20:01] jose: what project? [20:02] mhall119: lemme get a link for you, 1s [20:02] stokachu: yes 192.168.1.x/24 is private, 10.1.10.x/24 is [20:02] public [20:02] PCdude, and maas manages both dns/dhcp on 192.168.1.x? [20:03] stokachu: that is correct [20:03] PCdude, well im out of ideas, we know it's networking related [20:03] mhall119: https://code.launchpad.net/~jose/charms/trusty/couchbase/trunk the charm itself works perfectly, but automated testing is funky [20:04] thanks jose, I just want to read it [20:04] np, let me know if you have any questions [20:04] PCdude, im also inclined to say this is in maas territory [20:05] jose: just one question; "Do you want to learn how to make snaps?" :) [20:05] mhall119: maybe, I can give it a shot and if it's doable I can charm+snap [20:06] stokachu: yes network related and maybe LXD, I dont know. I have to say that both DNS and DHCP are working. I tried it both, by adding an random machine to it with linux already installed and tried to contact the internet [20:06] mhall119: live training in Europe? ;) [20:06] PCdude, well you can't deploy to containers on the machines [20:06] stokachu: ah ok, well at least we are a step closer to the solution [20:06] so i would say dhcp is not working [20:07] jose: dholbach will be doing a workshop, but we can do some bar-learning as well [20:07] mhall119: bar-learning sounds awesome, specially with a couple legal beers [20:07] stokachu: well if it is not working the nodes in the 192.168.1.x/24 domain should not even startup right? PXE is inherently connected to DHCP. [20:07] yay for legal beer! \o/ [20:08] PCdude, that's why i think you're existing your dhcp pool, maybe expanding the range didn't take effect [20:08] i dont know what doing an esxi snapshot/restore entails [20:09] jose: http://packages.couchbase.com/ gives me an access denied error, is that expected? [20:09] hmm, not at all. I'll check when I get back home. in class atm [20:10] oh, sorry to bother you in class (you should disconnect) [20:10] hehe no worries. it's just starting and things are getting set up, have a couple mins [20:11] looks like only partial URLs are being blocked, probably so you can't see a files list on the server, so no concern [20:11] stokachu: hmm maybe. I dont know about vmware, I think they just take a copy of the disk and put it back. [20:11] beisner: ping [20:12] stokachu: I think that is not the problem, but I could start over from the start, but that would take to much time rn [20:12] beisner: do you all deploy on kvm/nova on ppc64el? [20:12] PCdude, ok well you've got a good place to start debugging [20:13] ~ [20:13] ~ [20:13] stokachu: yeah indeed, at least I really wanna thank u! One step closer to an working openstack install :) === natefinch is now known as natefinch-afk [21:19] marcoceppi: out of interest [21:19] whats the process to get an interface listed on interfaces.juju.solutions [21:19] magicaltrout: yes? [21:19] i'm working on a drill one for jdbc connections etc so it would be good to allow other stuff to use it [21:25] magicaltrout: log in with LP, then add then click the + next to interface or layer [21:26] fill out the form, and submit ;) [21:26] ah the tool i try and avoid [21:26] will do [21:36] rick_h_: marcoceppi what was that bug number about xenial and lxd failing to create containers? [21:36] with juju [21:37] stokachu: it's a cloud-init issue, i think [21:37] - https://bugs.launchpad.net/ubuntu/+source/snapd/+bug/1621229 [21:37] - https://bugs.launchpad.net/ubuntu/+source/cloud-init/+bug/1576692 [21:37] Bug #1621229: snap upgrade to 2.14.2~16.04 in xenial lxc hangs [21:37] Bug #1576692: fully support package installation in systemd

[21:37] stokachu: sec, it's in that juju list email. /me goes to look [21:41] thanks [21:43] lazyPower oh fountain of useful code snippets [21:44] i need to write an interface, that when a user connects to drill the interface then goes and returns the ip addresses of the zookeeper units it has attached [21:44] how do i find out the ZK units? [21:50] aww [22:07] magicaltrout: o/ [22:07] Guess I'm not good enough? [22:16] you suffice marcoceppi [22:16] i just bug lazypower because he tends to have this stuff to hand ;) [22:17] magicaltrout - sorry about that, flakey internet here at the sprint. [22:18] lol no probs i'm just looking for a pointer [22:18] magicaltrout - as best as i can tell, you would need to coordinate with the consuming layer... cory_fu may know better, but from what i know, thats up to the implementation on which relation its scoped to, et-al. [22:18] i dont think there's a more deterministic approach to take there off the top of my head [22:20] okay so in my main drillbit charm [22:20] i keep a data_changed function which is a list of zookeepers [22:20] can my interface hit that? [22:22] so long as its your drillbit layer thats communicating to the interface, i think that keeps the levels of encapsulation [22:22] well i wrote the charm, now i'm writing the interface [22:22] so i think i'm good on those counts [22:57] i dont like interfaces they're like a blackhole of stuff i don't understand