[03:58] erm, am I right in my assumption that 1.25 doesn't have network-get, and that charmhelpers' network_get function is broken on it? [04:05] it seems that I am === frankban|afk is now known as frankban === skay is now known as Guest48273 [12:19] is it possible to get stdout from `juju run` in a streaming fashion? It seems that stdout from the `juju run` is only passed back to the caller after command completion, unless I'm missing something? === verterok` is now known as verterok [12:46] beisner: unless it would be OK to _not_ have 0-downtime, I'm looking at rolling style, but I could just tell it to do the release upgrade on all units in the model [12:46] :-/ sorry, wrong channel === Guest48273 is now known as skay [13:28] icey: If you don't need to run under a hook context, 'juju ssh' might be what you need. [13:28] stub: I suppose I don't need the hook context, I'm doing a `do-release-upgrade` ;-) [13:29] * stub wonders if he should add that as an action to the apt layer [13:33] could be nice, but I don't have layers ;-) [15:12] I'm trying to add Kibana to a machine in my model, but I get this error: "Error placing unit: unable to place a trusty unit on the xenial machine new8" [15:13] I'm unsure of the cause though [15:17] SuneK: Sounds like it's trying to deploy the trusty version of the kibana charm to a xenial machine. Try adding --series xenial to your juju deploy command [15:19] Makes sense, thanks [15:24] SuneK: np. Normally I'd expect juju to figure that out. Are you doing juju deploy kibana or deploying a bundle? Also what kind of cloud, lxd, manual? [15:25] zeestrat: SuneK yea, the issue there is that rather than assume it fails something it knows isn't valid [15:55] SuneK: if you're deploying a bundle, please let me know which one. i've made lots of changes around the elk stack in the last 2 weeks to polish things up for xenial, but i may have missed a bundle or two. [16:29] @kwmonroe I'm deploying the canonica-kubernetes-elastic bundle https://jujucharms.com/canonical-kubernetes-elastic/ [16:31] SuneK: roger that.. make sure you use the latest edge version to pick up xenial kibana: https://jujucharms.com/canonical-kubernetes-elastic/104. if you're doing this from cli, that's: juju deploy canonical-kubernetes-elastic --channel edge [16:33] kwmonroe: ok, I will do that, normally though I prefer stable versions for production use :-) [16:33] rick_h: So I built 2.1.2 with snapcraft which was really easy, but now I'm stumbling at no matching tools. I haven't messed much with uploading or customized agents so there's probably something obvious I'm missing. Am doing something dumb? https://paste.ubuntu.com/26541861/ [16:34] Got to go now though, i'll report back tomorrow [16:35] zeestrat: hmm, not sure tbh. I thought it would know this and auto upload from your own built version. [16:36] zeestrat: have to ask around, cory_fu any ideas? ^ [16:39] rick_h, zeestrat: Maybe 2.1.2 doesn't support bionic? You might try --bootstrap-series=xenial [16:40] cory_fu, rick_h That did the trick, thanks. Not sure why it tried to bootstrap bionic. [16:41] Is that default expected behavior? [16:41] zeestrat: I think it defaults to whatever your host system is. I assume you're running bionic? [16:41] Nope, xenial [16:41] Oh, strange [16:41] Maybe bionic is the default in simplestreams? [16:41] That something related to https://bugs.launchpad.net/juju/+bug/1727355 ? [16:41] Bug #1727355: Juju attempts to bootstrap bionic by default

[16:41]

[16:43] Yeah, that seems like it [16:47] cory_fu: Thanks for the snapcraft tip yesterday btw, that worked like charm (pun intended). [16:51] heh. Glad it worked out [17:02] zeestrat: cory_fu oh, didn't think of bionic issues === frankban is now known as frankban|afk [23:55] hi [23:55] anyone can help me with this : https://paste.ubuntu.com/26543884/ [23:56] is it related to this : https://bugs.launchpad.net/charm-ceph-osd/+bug/1731639 [23:56] Bug #1731639: get_partition_list fails