[01:19] Hi, trying to juju bootstrap to openstack private cloud... [01:19] I get this error, ERROR cmd supercommand.go:458 failed to bootstrap model: cannot package bootstrap agent binary: no prepackaged agent available and no jujud binary can be found [01:20] Any ideas how to move forward? Juju bootstrap is looking for jujud inside my local juju bin directory and not finding it, but cannot find any info on jujud. [01:22] som more debug: https://paste.pound-python.org/show/fNCbbes8Fv2Ul6wQx6rs/ [06:26] Hi all, I want to know about juju charm deployment on openstack instance [06:26] please someone help [08:08] anyone have a working RackSpace openstack Kilo juju controller setup? I keep getting credentials not found and no credentials needed for that cloud [08:08] any pointers or example yaml files appreciated for juju 2.x [08:10] Good morning Juju world [08:11] cacasmacas: you might have more luck asking at #openstack-charms [08:12] hi surf. What exactly you want to know? [09:03] hi here, if I deployed the canonical-kubernetes bundle charms with only private ethernet interfaces on my kubernetes-worker and if I add after this deployment public interface, wil Juju do something? does it ever need to? [09:04] s/wil/will/ [09:14] om2: you're missing simplestream metadata service for your cloud [11:03] BlackDex: We're running into the same issues when deploying OpenStack HA services on LXD on MAAS nodes. The nodes are pretty beefy, but deploying them all at the same time really bogs it down. [11:04] yea, i now told the bundle to use hacluster with a count of 3, and all the HA services like cinder, glance etc... i just deploy one unit. And afterwards i just do `juju deploy -n2 cinder --to lxd:1,lxd:2` [11:05] the cluster won't start untill there are 3, so this works for now [11:06] but if i'm correct, the juju-deployer package waited before atleast one was deployed, and after that it added the remaining units for that service/app [11:07] I came straight into juju 2.x so I haven't really tried juju-deployer. [11:11] It might be worth sending a mail to the mailing list to explain the issue and hear if anyone has some suggestions on deployment or performance improvements in Juju [11:47] I saw this clearwater+zabbix demo about autoscaling, is there any hint how that was achieved? [11:47] juju client + user installed to zabbix server and then using shell commands to scale the application? [11:48] I'm interested whether there's some "approved" way to implement autoscaling features? [11:49] zeestrat: where is your bootstrap located? Also an LXD? [12:32] Hi, should auto_peers for ntp be true, if we're going to use ntpmaster? [12:32] auto_peers config in ntp charm* [12:33] if we're *not* going to use ntpmaster** [12:33] hi all, how to boostrap juju on openstck cloud? [12:36] surf: you should check https://jujucharms.com/docs/2.0/help-openstack [12:37] I have created a guide over this (not a clean one though :) ), http://paste.ubuntu.com/23833150/ [13:35] surf, Did you mean how to deploy a workload using juju on top of openstack ? [13:47] deanman: yes i got the way to do. [13:48] there is couple things to be extra careful :) [13:49] ok great, i did it recently on Mitaka and could help if you want but i think junaidali's snippet is straightforward :-) [17:24] bdx - got your doc comment, and briefly glanced at the reactive code linked. I'd need to dive deeper in this context with a deploy to see if we cant do it another way [17:24] but it looks very similar to what we were referencing lastnight === wililupy_ is now known as wililupy [20:08] junaidali: I've changed the readme recently to hopefully explain this better; see https://git.launchpad.net/ntp-charm/tree/README.md for latest [20:09] junaidali: I have to run now, but feel free to drop questions here and I'll do my best to answer.