=== thumper is now known as thumper-headdown === thumper-headdown is now known as thumper [08:00] Good morning Juju world! [09:16] kjackal: arent you a happy bunny [09:17] admcleod: what's up crazy boy? [09:34] morning [09:34] last day of the week, lets see if i can get somehting working [09:37] kjackal: crrrrrazyboy [09:37] cnf: fingers crossed! [09:37] hmm [09:38] cnf: seeing your frustration makes me want it to work too [09:38] admcleod: thanks :P [09:38] cnf: as the americans would say, we're "rooting for you". im not american though so thats pretty weird [09:39] so, adding constraints to my stuff broke everything [09:39] but i need to, or my lxd containers don't have networking [09:40] or, add a single static route to the controller? [09:41] ah you have no default gateways right. ok. [09:44] right [09:44] and adding a route means extending the vlans to a different point in the network [09:45] i wonder if you could add the controller to all the spaces [09:45] well, no, afaik juju advertises the controller ip [09:46] so if i add constraints: "spaces=space-maas" to an application [09:46] maas no longer offers [09:46] machines [09:48] hmm [09:54] cnf: remind me again why the applications in lxd containers arent bound to space-openstack-mgmt? [09:54] they are? [09:54] they aren't to space-maas [09:54] oh, im looking at a status paste from yesterday rather than a bundle [09:54] ok [09:55] hmm, and now maas is not generating logs o,O [09:55] why did maas stop logging? o,O [10:07] ok, that's fixed, i think [10:19] ok, i'm deploying something [10:19] we'll soon see what :P [10:20] heh [12:06] morning all. I've got an issue bootstrapping juju to local LXD https://pastebin.canonical.com/183637/ it's building the container but then fails towards the end, saying it can't connect but the bridge is up and I can access the port via telnet [12:22] joedborg: did you try https://bugs.launchpad.net/juju/+bug/1633788/comments/14 ? [12:22] Bug #1633788: juju 2.0.0 bootstrap to lxd fails (connect to wrong "remote" IP address)

[12:24] joedborg: the last part atleast [12:25] admcleod: I can try the dev version and see [12:25] joedborg: more just the yaml mod [12:25] I did try the cloud def, but not with dev version [12:26] joedborg: what ver do you have instlled now? [12:27] admcleod: I had 2.1.2, now 2.2-a [12:29] admcleod: get the same issue [12:30] joedborg: presumably you've named the cloud in clouds.yaml the same as the name you're using to bootstrap [12:31] admcleod: yeah, so i just copied "lxd" name, can see that in the cloud list and then bootstraped with "lxd" [12:32] joedborg: and logs still show it trying to connect to .1 instead of controller ip from start of bootstrap? [12:33] joedborg: it was always trying to conenct to the right ip [12:33] admcleod: 2017-03-24 12:32:20 ERROR cmd supercommand.go:458 new environ: Get https://10.201.51.1:8443/1.0: Unable to connect to: 10.201.51.1:8443 [12:34] admcleod: ah, i see what you mean [12:34] joedborg: i thought your controller was .123? [12:34] right [12:35] admcleod: this is the IP of LXD [12:35] admcleod: the bridge connector [12:36] joedborg: oh yeah. :) [12:37] admcleod: which i can access and port 8443 is open [12:37] admcleod: so I think what ever's on that port isn't replying as juju expects [12:37] joedborg: so as beisner asked, proxy set anywhere? lxc config? [12:38] mm [12:38] https://www.irccloud.com/pastebin/aK1Xhdmk/ [12:38] what ver lxd? [12:38] 2.12 [12:38] admcleod: it updated itself recently [12:39] * admcleod updates [12:39] joedborg: what repo did you get 2.12 from? [12:40] admcleod: xenial repos i believe [12:40] admcleod: now checking, but don't think i have any PPAs [12:41] admcleod: I tell a lie [12:41] admcleod: deb-src http://ppa.launchpad.net/ubuntu-lxc/lxd-stable/ubuntu xenial main [12:46] joedborg: im confused, 2.12 isnt an lxd version i can see anywhere, nor lxc 2.12.. [12:47] admcleod: $ lxd --version [12:47] 2.12 [12:47] $ lxc version [12:47] 2.12 [12:49] joedborg: apt-cache policy lxd ? [12:51] https://www.irccloud.com/pastebin/M0nDXowp/ [12:51] admcleod: ^^ [12:55] joedborg: ..installing... [12:56] admcleod: ta [12:57] joedborg: bootstrapping [13:00] admcleod: did lxd launch work for you post updatE? [13:02] joedborg: lxd? launch? lxc launch yes.. [13:03] joedborg: back in 15 [13:03] admcleod: yeah i meant that. that didn't work for me until i purged and reninstaled [13:03] admcleod: (y) [13:16] joedborg: bootstrap OK [13:22] hi guys [13:22] I've a question on how to rename a charm which has already been pushed ? [13:22] In a same way, is it possible to remove a charm ? [13:23] admcleod: with juju 2.1.2? [13:24] yes [13:25] joedborg: yep [13:26] gabriel: I think rick_h mentioned that the remove API is under construction. Search for "remove a charm" in https://irclogs.ubuntu.com/2017/03/23/%23juju.html [13:27] zeestrat: gabriel yes, the delete API is in testing and not in the cli for end users yet. [13:27] gabriel: just repush it to the new name and file a bug for the delete and I'll get someone to peek at it [13:28] thanks! [13:28] joedborg: https://stgraber.org/2016/04/18/lxd-api-direct-interaction/ [13:31] joedborg: just curl --insecure https://10.201.51.1:8443/1.0 should be enough to see if the api is reachable /functional from the controller [13:32] admcleod: let me check [13:36] rick_h: I just opened a support request to remove the old charm [13:36] gabriel: what's the link and I'll ping someone === disposable3 is now known as disposable2 [13:44] rick_h: thank you very much, but I cannot find the support ticket link [13:44] gabriel: oic, that support link. Ok [13:44] gabriel: yea that goes through a new system for the jaas beta stuff. I'll have to see where that ends up. Thanks for the heads up [13:45] I see, thanks ! :) [13:56] admcleod: yeah that curl works fine [13:58] joedborg: i figured, since it looks more like an network error in your paste [13:59] admcleod: but i don't get what kind of problem because all the stuff it's trying to access is there [14:00] joedborg: *nod* [14:06] rick_h: thanks a lot, it looks removed ! [14:15] gabriel: oh cool, that was fast [14:17] ok [14:17] next problem :/ === admcleod is now known as admcleod_afk [14:56] next challange you mean ;) [15:27] admcleod: just purged lxd and juju again (with the bridge and zfs pool), reinstalled with snappy and get the same issue. so either it's a bug that only surfaces on my laptop's setup or there's some configs somewhere not being purged [15:28] joedborg: did you lxd attach to the container to check the connectivity to the controller? (lost my scrollback) [15:28] joedborg: s/lxd/lcx [15:29] lxc [15:32] admcleod: I can still curl it yeah [15:34] joedborg: weird [15:35] admcleod: yeah i suspect purging isn't removing a config file somewhere [16:04] mthaddon: would it make sense for apply-config-changes in mojo to check for template files that have changed? [16:11] hey lazyPower am I suppose to be able to us S3 storage out of the box with CDK regardless of infra? [16:12] lazyPower I am aware of having to setup a storage class, etc. but otherwise it should just work, right? [16:20] Budgie^Smore: I think that depends on IAM profiles being assigned to the ec2 units correct? [16:21] lazyPower don't think so, cause what if I wanted to build my own bare metal backed by s3 storage, ec2 doesn't come into it [16:21] so this backed by s3 is whats confusing me [16:21] s3 isn't listed as a supported volume type [16:21] in what context are you using s3 sotrage [16:21] *Storage [16:22] lazyPower does there need to be IAM policies, roles, other access issues overcome - probably but nothing to do with ec2... persistent volumes [16:23] Budgie^Smore: so if you're talking persistent volumes, you're referring to things like EBS, GCEPD, CinderVolumes, CephRBD, or gluster/nfs shares [16:23] last i checked, object stores were not viable backend volume types [16:23] lazyPower EBS falls under S3 [16:24] kinda [16:25] not really [16:25] EBS is a block device [16:25] s3 is an object store [16:27] EBS snapshots are stored in S3. so if you use `ec2-create-snapshot`, you are copying your EBS to S3. [16:29] OK right now that is a semantics arguement at worst... I should be able to use EBS PV including using PVC to create it running on bare-metal [16:29] right? [19:50] how can i add ssh key to machines [19:50] a new one [19:50] i added the key to the model [19:51] but i want it to distribute it to all machines and lxd containers [19:51] ybaumy: have to ssh-copy-id them to the machines and containers. ybaumy or do a juju run of 'ssh-import-id' maybe [19:52] rick_h: ok that works [19:52] thanks [22:03] marcoceppi: https://github.com/juju-solutions/layer-etcd/pull/85 [23:28] hey guys, what're some things you don't like about juju and what can be improved? [23:35] i have a list... [23:36] well let's see it [23:47] https://bugs.launchpad.net/juju