[01:00] babbageclunk: here's a small juju/os PR https://github.com/juju/os/pull/17 [01:01] wallyworld: ok, taking a look [01:10] babbageclunk: ty, disco support runs out on 18th jan [01:10] oh right [01:21] babbageclunk: and here's the tiny juju dep update https://github.com/juju/juju/pull/11068 [01:22] wallyworld: no Gopkg.lock change? [01:33] wallyworld: approved [01:34] babbageclunk: yeah, i forgot to push it, sorry [01:43] nw [01:56] babbageclunk: hopefully last one, https://github.com/CanonicalLtd/juju-qa-jenkins/pull/356 [01:57] wallyworld: lgtm! [01:57] ta [01:57] now to generate jobs [05:24] tlm: can i get a +1 on a unit test fix? https://github.com/juju/juju/pull/11070 [05:28] yep [05:30] done, maps have order :| [05:33] great ty [14:47] achilleasa, can you CR https://github.com/juju/cmd/pull/74 [14:48] stickupkid: I will take a look in about 20min if that's OK [14:48] achilleasa, sure nps [14:55] Hi guys, I'm trying to bootstrap OSM via juju but am getting this error: ERROR failed to bootstrap model: cannot start bootstrap instance in availability zone "my-juju-machine": no matching image found [14:57] Command used: sudo juju bootstrap localhost osm [14:58] Harm133: hmm, is there something setup in the .local/share/juju/clouds.yaml that has that "my-juju-machine" in there? [14:58] There is no clouds.yaml file in that folder [15:00] hmm, I deleted the folder [15:00] Now I have something I can work with, a timeout error regarding the clouds.syaml file [15:09] Harm133, example of the error ? [15:09] Hmm, I'm getting this now : [15:09] ERROR failed to bootstrap model: cannot start bootstrap instance in availability zone "my-juju-machine": Failed remote image download: Get https://cloud-images.ubuntu.com/releases//streams/v1/index.json: Unable to connect to: cloud-images.ubuntu.com:443 [15:09] while a curl towards that URL works fine [15:10] Harm133: hmm, so that's probably coming from the lxd container that's started up so it might be an issue with lxd containers reaching the outside world? [15:10] that could be, can I pass a proxy towards the container? [15:11] typically you'd adjust the bridge on the host I believe. There are proxy settings you can set in bootstrap config if you need to for https and such [15:12] I've altered the env on the host but that does not seem to work, how would one go about configuring proxy settings within the bootstrap config? [15:17] Got it, lxc config set core.proxy_https / lxc config set core.proxy_http [16:28] stickupkid: you’re too fast! :-D. [16:28] stickupkid: didn’t get a chance to link my results. ha [16:28] stickupkid: ty! [16:39] hml, was waiting for a thing to land :D [16:39] hehe [16:39] stickupkid: ha! [17:39] wow, if you don't have a credentials yaml file, all hell breaks loose in the CLI [18:01] stickupkid_: yea...hitting that with these demos [18:02] and if you don't have the cloud forget credentials [18:13] hml, if you get a chance, no rush though - https://github.com/juju/juju/pull/11072 [18:13] stickupkid_: ack [19:02] hello, happy new year! [19:03] does anyone know if there is a way to migrate a model into jaas from a non-jaas/local auth juju controller [19:04] ? [19:04] guessing this probably can't be done - due to the users of the model needing to match in the source and target controllers [19:07] bdx: yea, it can't be done but more because I don't think JAAS supports model migrations due to it thinking it's one controller vs multiple [19:07] ok [19:07] bdx: it might be doable by a JAAS admin user if they had full access to the local model, but would be a test [19:09] bdx: what's a good bundle that shows off a service with different bits to it that's not a platform like openstack, k8s? [19:11] https://paste.ubuntu.com/p/zJ5VQ49446/ [19:11] bdx: hah good stuff [19:14] killed 2/3 juju controllers in aws yesterday after having recreated the infrastructure in jaas [19:14] yay? [19:14] one of them was the longest lived instance in our aws account [19:14] yeah [19:14] nice [19:15] yeah - no reason other then to cut down on our operational costs [19:15] yea, that's good stuff though [19:15] totally [19:16] this graylog model on this controller cant be spun down and recreated as easily as the others [19:16] so I was hoping there might be a way to migrate it into jaas so I could finsh off 3/3 :) [19:16] yea, I wish but not quite [19:17] not a big thing ... but if someone is down to poke at it for a few minutes I can create a sample model for them [19:17] totally [21:34] babbageclunk: https://github.com/CanonicalLtd/juju-qa-jenkins/pull/359 pls? nice and short [21:38] hml: looking [21:40] hml: approved [21:52] babbageclunk: ty [22:12] wallyworld: You around? [22:12] Or kelvinliu? [22:13] hi cory_fu: [22:14] how can i help u [22:14] kelvinliu: I'm getting an unexpected error on a CMR with a k8s charm and I was hoping you could help me understand what's going on. [22:15] kelvinliu: https://pastebin.ubuntu.com/p/7wHvXy3Nf4/ [22:15] That works fine when not using a cross-model relation [22:16] I'm also surprised that the unit name is not replaced by the UUID like it is in non-k8s charms. [22:19] cory_fu: mind share the bindle or how did u relate the cmr [22:20] kelvinliu: Well, I hit this while testing https://github.com/johnsca/charm-gitlab-k8s using the new framework, along with cs:~charmed-osm/mariadb-k8s. For the CMR case, I just did a separate model in the same k8s cloud and did a normal offer / consume. [22:21] SAAS Status Store URL [22:21] mariadb-k8s active microk8s admin/other.mariadb-k8s [22:22] ok thanks [22:25] I will take a look and let you know i found. ty cory_fu [22:26] Ok, thanks [22:28] np [22:36] cory_fu: i would have thought it would give the uuid also. there's no difference in k8s vs cloud in how cmr works so it does seem weird. it maybe well be a latent bug