[07:37] <gaurangt> Hi, I'm using MAAS in KVM virtual environment. When I try to deploy the VMs with xenial image, the cloud-init job tries to update the kernel and it is very very slow. Any idea on what might be going wrong?
[08:02] <blahdeblah> gaurangt: More than likely it's a slow connection between you and archive.ubuntu.com/security.ubuntu.com; you can specify a local mirror in the MAAS settings.
[08:03] <gaurangt> yeah.. I've tried that as well. I guess the problem is not with downloading packages. Unpacking and installing them is very slow.
[08:04] <gaurangt> blahdeblah: http://paste.ubuntu.com/23445514/
[08:04] <gaurangt> this takes a long time
[08:17] <gaurangt> Is there any way to increase deployment timeout in MAAS?
[08:18] <gaurangt> one other observation - The problem occurs when I deploy more machines. Timeout doesn't occur for a single machine deployment.
[08:18] <blahdeblah> gaurangt: Other ways you can speed it up are using bcache, and (if you're just using the KVM machines for testing and don't care if they break) using cache=unsafe for your VMs
[08:19] <gaurangt> where should I specify these settings?
[08:19] <blahdeblah> bcache is done when you install the bare metal machine - you need an SSD to cache your spinning rust HDs
[08:20] <blahdeblah> cache=unsafe is done when you run virt-install for the VM
[08:20] <gaurangt> ok
[08:21] <gaurangt> I'm using virt-manager for creating the VMs
[08:21] <gaurangt> i have set cache option as none
[08:21] <blahdeblah> I have a rather old core 2 duo machine I use for testing, and it's quite speedy enough when I put both bcache and cache=unsafe
[08:22] <gaurangt> the other value options are default/writeback/writethrough
[12:19] <mup> Bug #1640147 opened: [2.1] rsyslog is flooded with "no link-local IPv6 address for $IFACE" on commissioning <MAAS:New> <https://launchpad.net/bugs/1640147>
[15:28] <mup> Bug #1640201 opened: [API] node parameter silently ignored when creating DHCP snippets if global_snippet set to True <maas-cli> <MAAS:New> <https://launchpad.net/bugs/1640201>
[15:59] <Braven> Hello
[16:01] <Braven> Can I have maas manage dhcp but set the scope to 0 host available. But then allocate static IP
[17:58] <mup> Bug #1640259 opened: [2.1 UI] IP Ranges table no longer showing in Subnet Details page. <MAAS:In Progress by newell-jensen> <https://launchpad.net/bugs/1640259>
[19:52] <roaksoax> /win/win 6
[20:44] <mup> Bug #1640298 opened: multiple subnets dhcp shared-network does not work with dhcp-relay <MAAS:New> <https://launchpad.net/bugs/1640298>
[20:59] <nturner> Anyone want more info about the state of https://bugs.launchpad.net/maas/+bug/1640300 before I restart maas?
[21:02] <roaksoax> nturner: the postgres logs
[21:04] <nturner> roaksoax: ok, I can attach those
[21:04] <roaksoax> nturner: thanks
[21:07] <nturner> ok, attached
[21:07] <roaksoax> nturner: ty
[21:09] <roaksoax> nturner: approximately, how many machines do you have ?
[21:09] <nturner> 36
[21:10] <roaksoax> nturner: yeah that sounds like something is leaking
[21:10] <nturner> want any other logs?
[21:10] <roaksoax> nturner: i think we have an idea of what it may be. So postgres logs + all maas's should be enough
[21:11] <nturner> ok, I'll tar up and attach /var/log/maas too
[21:11] <roaksoax> nturner: ty
[21:14] <mup> Bug #1640300 opened: django.db.utils.OperationalError: FATAL:  remaining connection slots are reserved for non-replication superuser connections <MAAS:New> <https://launchpad.net/bugs/1640300>
[21:14] <mup> Bug #1640301 opened: Deploying servers with multiple disks fails <MAAS:New> <https://launchpad.net/bugs/1640301>
[23:08] <mup> Bug #1640333 opened: No rack controller available when trying to enable DHCP on subnet on tagged VLAN <MAAS:New> <https://launchpad.net/bugs/1640333>