=== scuttle|afk is now known as scuttlemonkey === scuttlemonkey is now known as scuttle|afk === med_ is now known as Guest6511 [07:47] Good morning Juju World! === babbageclunk is now known as babbageclunk-afk [13:38] rick_h: you blew up the JAAS hack! [13:38] :) [13:38] bdx: who what me? [13:39] - ha [13:41] bdx: it was a cool hack and had me finding out Juju didn't work the way I thought it did. You *blew* my mind dude [13:43] +1 [13:43] it is great to have that better understanding now. [13:44] hehe, nice! yea!! === jamescomey is now known as jamescomey007 [15:07] is there a different channel for Charm - as in charm build command or is this the group [15:10] I am having issues bootstrapping a vm based controller on MAAS. Is this the right place to ask questions? [15:24] I'm having a problem with CHARM Build --series centos7 - which brings in python files written for ubuntu. Who should I talk to? [15:46] jac_cplane: you're in the right place [15:50] jac_cplane: https://lists.ubuntu.com/archives/juju/2017-May/thread.html - search for centos on that page [15:51] jac_cplane: specifically, this one might help: https://lists.ubuntu.com/archives/juju/2017-May/009053.html [15:56] Yes - i went through this. and We already have a charm that deploys perfectly with centos7. the problem is when I try to get the charm ready for the charm store. Charm push . or charm build adds basic.py which includes apt-get commands [15:56] jcromer: would you mind adding your bootstrap debug pastebin here. We'll figure out who we can get to help shortly [15:57] thedac: Sure [15:57] https://pastebin.com/Hd89pz4E [15:57] rick_h: who on the juju team can we get to help out with a bootstrap problem for jcromer? ^^^ [15:58] thedac: /me reads backlog sec [15:59] rick_h: the bootstrap debug just stops at fetching tools. No error. [16:00] thedac: I see, I assume that's hung there for a while? Is the instance coming up in maas and can it reach streams for the agent? hmm, why are we using 2.0.2? [16:01] thedac: jcromer I'd suggest poking at the log it mentions there: Logging to /var/log/cloud-init-output.log on the bootstrap machine [16:01] and see if anything is there [16:01] jac_cplane: https://github.com/lenovo/workload-solution/wiki/python2-charm#basicpy documents how fengxia worked around that [16:02] thedac: jcromer and 2.0.2 is fishy as 2.0.4 is latest on the 2.0.x series and we should at least be on 2.1.3 for anything coming up unless we're testing something old? [16:02] rick_h: long story on the 2.0.2. But the bootstrap node is in a broken state so he can't get to that log [16:02] thedac: he can't ssh to it directly? [16:02] thedac: i can update juju [16:03] i am in [16:03] thedac: jcromer so I'm not sure. The machine came up enough for cloud-init to kick off and so if maas is setup with keys I'd expect to be able to connect with those [16:04] sorry, had been attempting to use juju ssh [16:04] direct ssh works [16:04] thedac: jcromer and I'd see if there's something in the logs, I'd think there's something in the network going out [16:04] thedac: jcromer which is causing it to fail. [16:04] so what version of juju do you recommend that i update to? [16:04] jcromer: right, so getting the agents is getting the binary that is juju so until that's done you're not in juju land [16:04] jcromer: 2.1.3 is the latest stable release [16:05] jcromer: if you have to be on 2.0.x per thedac's long story then 2.0.4 so it has the security fix that was released a weekish ago [16:05] i don't have to stay on 2.0.x [16:05] jcromer: k, then yea 2.1.3 and from here I'd debug why it isn't getting the agent download via sitting on the machine provisioned and checking out if it can reach streams and if there's anything in the cloud-init log there [16:07] ok, i will update juju, attempt bootstrap again and then check logs if i still see issues [16:07] thanks, for the help and information [16:07] jcromer: definitely, unfortunately a lot of the team is traveling and hitting weekend time so it might be a little bit quiet but I'll try to keep an eye out for your next steps [16:08] thanks, greatly appreciated [16:08] jcromer: I strongly assume that there's something like an apt proxy sitting there that needs to be setup for ssh, or an agress network filter, or something fishy causing the cloud-init script to hang trying to get the juju binary from the url it's telling you in the logs there. [16:09] but that's just my own experiences I seem to hit when I get something that looks like this. [16:22] - ok -- the last work-around looks promising. do you know if this be interfered with if I do a Charm push cs: ? [16:22] will layer/basic gets overr written when put into the charm storr [16:38] jac_cplane: charm push should work fine [16:39] jac_cplane: you would want to charm build, then make local edits to basic.py, then push [18:57] o/ juju world [18:58] \o Budgie^Smore [18:58] whats news budge [18:58] hating on bash exit 1 status code right now [18:59] for some reason I can't get my preseed file to echo a line into a new file :-/ [19:04] booger :\ [19:09] you mean like echo "a line" > a_new_file ? [19:20] jrwren yeah [19:24] I am putting in-target in front of the line too [19:40] Hey guys quick question. I'm deplyoing the openstack base bundle via juju onto a maas set of managed instances. Do I need to manage the dhcp of the servers and containers themselves from maas or can juju manage this? Or should I have something else managing DHCP for all those networks? Thanks in advance! === mup_ is now known as mup