[07:47] <stub> tvansteenburgh: I got another lxc failure from CI, same as last time - ERROR there was an issue examining the environment: cannot use 37017 as state port, already in use
[07:48] <stub> (there had been successful runs in the interim, so it seemed to survive ok for a few days).
[09:12] <vahid> Hi. We have a juju environment (manual provisioning version 1.20.4) with over 140 machines and several services on each one.
[09:12] <vahid> We are using this for over 9 month and now it's getting completely unstable.
[09:13] <vahid> Problem started after I remove a subordinate service and juju-agents crashed on over a thirds of machines (about 43 of them) and they didn't remove and keep reporting error:
[09:13] <vahid> "ERROR juju.worker.uniter.filter filter.go:116 tomb: dying"
[09:13] <vahid> "panic: runtime error: invalid memory address or nil pointer dereference"
[09:13] <vahid> After few days juju agent on machine-0 became completely unstable and it goes down after few minutes, so we couldn't even get status of the system.
[09:13] <vahid> We decided to upgrade juju so we blocked all machines api request to machine-0 with iptables to make stable and then run upgrade-juju command.
[09:13] <vahid> But after that we encounter another bug (again!):
[09:13] <vahid> "some agents have not upgraded to the current environment version 1.20.14: machine-803"
[09:13] <vahid> machine-803 have had a problem while adding and its agent state froze on pending. I tried to force remove this machine but the machine is still there
[09:16] <vahid> Any body knows what we can do in this situation?
[12:49] <vahid> Somehow we managed to run upgrade-juju with "1.22.8", command successfully finiched but we lost juju agent on machine-0!
[12:49] <vahid> Now we receive this error on machine-0:
[12:49] <vahid> exited "state": cannot log in to juju database as "machine-0": unauthorized mongo access: auth fails
[13:36] <icey> did juju deployer change to require sending deployment name (machines, relations, series, services) recently? Is there a way to deploy the bundle.yaml as specified otherwise?
[14:20] <icey> that moment when google can find a charm in the charm store but the search on the charmstore does not
[14:20] <icey> :-/
[14:35] <xnox> hello, I am using local provider, and then I did $ juju ssh 0
[14:35] <xnox> however i'm getting permission denied (publickey,password)
[14:35] <xnox> was i meant to setup an ssh key somewhere before hand?
[14:55] <xnox> are xenial charms published yet?
[15:08] <xnox> is juju ssh 0 supposed to work with juju local on wily?
[15:41] <asanjar> kwmonroe: hey
[16:02] <mbruzek> asanjar: You are back!
[16:06] <kwmonroe> hey asanjar!
[16:14] <jose> oh, hi asanjar!
[16:22] <lazypower> vahid thats troubling!
[16:22] <lazypower> vahid have you filed a bug report with what you've described above?
[18:09] <icey> so, juju won't let me provision on d2 instance types, says invalid type and gives me a list of valid values
[18:10] <lazypower> icey using the alpha?
[18:10] <icey> lazypower: 1.25.0-elcapitan-amd64
[18:11] <lazypower> icey https://bugs.launchpad.net/juju-core/+filebug
[18:13] <icey> lazypower: https://bugs.launchpad.net/juju-core/+bug/1535838
[18:13] <mup> Bug #1535838: Juju won't let me use some instance types on AWS <juju-core:New> <https://launchpad.net/bugs/1535838>
[18:13] <lazypower> thanks icey
[18:14] <lazypower> icey can you add juju version to the bug?
[18:14] <lazypower> i think there's work thats been done / is being done - to remove this static map and update with streams or something
[18:15] <icey> sure thing