=== lounge-user85 is now known as heckles1
[01:29] <lounge-user55> heckles1: welcome to #juju
=== lounge-user55 is now known as bdxbdxbdx
[01:30] <bdxbdxbdx> heckles1: you made it!
[02:02] <heckles1> that i did
[10:38] <ybaumy> moin
[11:27] <ybaumy> DocScrutinizer05: join us at #opensuse-de-chat
[16:47] <bdx> this may be a silly question, why can't JAAS support all the providers?
[16:48] <bdx> if your maas or openstack had public endpoints that JAAS could talk to, it would theoretically be possible right?
[16:48] <rick_h> bdx: in theory yes
[16:50] <bdx> rick_h: it probably isn't the best idea to have the agents talking to the controller over the wan though ... or is it?
[17:02] <rick_h> bdx: yea, not a good idea. Between networking/egress complexity, having every hook fire event going over the wan, etc.
[17:03] <bdx> yeah
[17:03] <bdx> thats what I was thinking
[17:03] <bdx> rick_h: how come `juju users` doesn't work in JAAS?
[17:04] <bdx> is there another way to get the users in my model?
[17:04] <rick_h> bdx: because it's SSO backed and doesn't list all users in SSO
[17:04] <rick_h> bdx: juju show-model xxx
[17:04] <bdx> oh niceee
[17:04] <bdx> rick_h: thx
[17:50] <bdx> rick_h: https://gist.github.com/jamesbeedy/eaca40c935ffc8c2cfaf0ca7e0f33d4d
[18:19] <fallenour_> So
[18:19] <fallenour_> who do I scream at for trying to force me to use ubuntu username only when I could ssh in as root?
[18:19] <fallenour_> because that jackass has locked me out of my entire juju environment, and cost me 8 days of work
[18:19] <fallenour_> Oh, and wiped out about 40 non profits entire infrastructure
[18:20] <fallenour_> So fuck you *insert name here*
[18:20] <fallenour_> Other than that, juju is fucking stellar
[18:21] <fallenour_> Also, how do I upload my custom build xenial openstack build out? Id like to share that with the community, thankfully I have most of my work in yamls, its just a bitch waiting for it all to build out
[18:23] <rick_h> fallenour_: not sure on the first one what you mean there. The ubuntu images used in cloud/lxd deployments and such are ubuntu which by default uses sudo and no root login. I'm guessing that conflicts with something you were expecting?
[18:25] <rick_h> and for the second one, upload your custom xenial openstack build out, is that the bundle? ... oic you want to snapshot the images themselves to use as the deployment. Because the charms allow things to be mutable and respond to events/their environment the tricks to use custom images takes a bit of hackery and aren't fully supported
[18:25] <rick_h> bdx: nice with the plugin :)
[19:00] <fallenour_> ughhh
[19:00] <fallenour_> so sad :(
[19:04] <fallenour_> anyone know how to create a permanent bundle for juju charms store?
[19:07] <rick_h> fallenour_: the easy way is to use the juju-gui export button to export the yaml. The you use the charm command to upload to the store. Let me get the docs link for you
[19:08] <fallenour_> I got the charm yaml ready, I just need the guide if one exists
[19:08] <fallenour_> shweeet!
[19:08] <rick_h> fallenour_: https://jujucharms.com/docs/2.0/charms-bundles
[19:09] <fallenour_> HAZAAH!
[19:09] <fallenour_> time to add a couple hundred charm bundles
[20:34] <fallenour_> seriously guys
[20:34] <fallenour_> whats the deal with juju
[20:34] <fallenour_> why does it need a machine for maas
[20:34] <fallenour_> a machine for juju
[20:34] <fallenour_> and then ANOTHER MACHINE for ANOTHER JUJU controller, when I already have juju installed?
[20:34] <fallenour_> why do I have to waste 3 servers for this shit?
[20:35] <fallenour_> conjure-up is seriously driving me insane
[20:38] <stokachu> Why
[20:38] <fallenour_> Why indeed @stokachu
[20:38] <fallenour_> why cant it all deploy and manage from a single server, and then the juju enable-ha simply clone itself via maas to another box, and build another regional maas and juju for HA?
[20:39] <fallenour_> why waste so much hardware on something that uses next to no resources whatsoever?
[20:39] <stokachu> Sorry I meant conjure-up
[20:39] <fallenour_> dhcp and IPMI pretty much anything over lan eats next to no ram, no cpu, and little bandwidth
[20:39] <fallenour_> yea conjure-up
[20:39] <fallenour_> and then it proceeds to grab anothe rbox for juju, a second one
[20:39] <fallenour_> for 3 wasted servers
[20:40] <fallenour_> and heaven forbid you are intereste din landscape, because that eats a 4th
[20:40] <stokachu> So there is a architect button that will let you combine apps on same machine
[20:40] <fallenour_> you might as well buy a damn rack just to test the stack
[20:40] <fallenour_> no
[20:40] <fallenour_> unfortunately not
[20:40] <fallenour_> at this rate, I might as well point my ukranian team at rebuilding all of canonical stack
[20:41] <fallenour_> im half tempted to ask them on twitter if they hired Microsoft Teams after the lay offs with all the hardware they are wasting
[20:41] <fallenour_> If I wanted to blow this much resources on wasted hardware, I would have installed vSphere from the get go with NetApp
[20:41] <stokachu> Well it is Sunday here and I'm working on some no related canonical stuff but will be willing to help later today
[20:41] <stokachu> Yea I get it you're upset
[20:41] <stokachu> I'll try to help as best I can
[20:41] <fallenour_> Its not that im upset, I can afford the wasted hardware
[20:42] <fallenour_> My issue isnt me, my issue is non-profits, and other less well off individuals and companies
[20:42] <fallenour_> they cant
[20:42] <stokachu> Right well we don't want you wasti g hardware
[20:42] <fallenour_> Is it possible to put maas and juju all on one system?
[20:42] <fallenour_> I know the official guide says 2 systems
[20:43] <stokachu> In a sense you can start a KVM instance on the maas machine
[20:43] <stokachu> Use that as the juju controller
[20:43] <fallenour_> but Ive monitored the resource util,a nd even on older Dell r610s, the util even with deploying an entire rack at a time is less than 50%
[20:43] <fallenour_> yea I figured that much, but why cant I simply just do conjure-up --boostrap-to localhost ?
[20:43] <fallenour_> and install conjure-up on maas?
[20:44] <stokachu> Just the way juju works
[20:44] <fallenour_> if its going to do a loopback to a local host, itll just hit itself at http://localhost:5240/MAAS instead
[20:44] <fallenour_> I know thats how juju works that way, but why?
[20:44] <stokachu> I can help you more later when I'm back in the office tonight
[20:44] <stokachu> Just the way juju works
[20:44] <fallenour_> was there ever an rfc done for it?
[20:45] <stokachu> Don't think so
[20:45] <fallenour_> shame. I wouldnt mind helping contribute code cycles to get it to work, I know it can, we all do, but I feel theres something else that wasnt released or noted thats the reason why. Do you have any ideas?
[20:46] <stokachu> Sorry I don't, but if you give me a little time today I'll be back in the office to assist more
[21:03] <fallenour_> I think i just hit the jackpot of all ideas
[21:03] <fallenour_> o.o
[21:03] <fallenour_> Ok so stay with me
[21:03] <fallenour_> LXD container autodeployed with MAAS, with a juju container
[21:03] <fallenour_> python installer to auto deploy a ha pair to another box via juju via maas, localhost > remote host
[21:04] <fallenour_> it deploys HA, then builds out openstack standard via conjure-up + relevant options
[21:05] <fallenour_> minimal changes to the config guides as well