[12:32] Bug #1762344 changed: JS tests random failure with "test caused a page reload" error [16:45] Bug #1786736 changed: curtin fails to open/erase disk with existing MDADM RAID configuration when deploying Windows images only [18:32] roaksoax: per https://bugs.launchpad.net/maas/+bug/1786567, what/where are the storage logs you are referencing? [18:35] oohh, I think I found the duplicate possibly https://bugs.launchpad.net/curtin/+bug/1770280 [18:35] my bad, sorry for the noise [18:36] roaksoax: could you advise what the best way for me to pick up the changes introduced to fix 1770280 [18:37] I'm in the middle of trying to get this deploy out the door and am blocked by this [18:38] possibly I should try getting the node into recovery mode, ssh in and try wiping the disks by hand [18:38] s/recovery/rescue/ [18:41] I think I may be hitting a different bug now putting the node into rescue mode [18:43] https://bugs.launchpad.net/maas/+bug/1787006 [18:48] Bug #1787006 opened: node fails to enter rescue mode [18:50] If I have 2x physical machines and want to install MAAS+Openstack - I'm assuming I would choose 1 physical machine as the MAAS host/controller (including openstack install) and the 2nd machine would be added as an empty bare metal chassis via pxe? [19:03] dskull: with only 2 machines, you may want to checkout `conjure-up openstack-novalxd` [19:04] woudln't that only be a single server install? [19:08] oh i see, its either lxd or kvm [19:08] dskull: yeah, that is essentially what you are doing if you have 2 nodes and make 1 of them your maas controller + juju controller and the other one you deploy openstack on [19:08] its just alot of extra/unneeded things if you only have 2 nodes, but it can totally be done [19:09] conjure-up openstack-novalxd - will probably be an easier way to start if you are just digging into things trying to get a POC [19:10] run it on one of those machines [19:10] so i'd essentially have a MAAS controller running with an extra 60gb of ram that isn't utalized? I'd kinda want MAAS+openstack to run off the same physical server and expand to the 2nd when necessary - i supposed i'd have to virtualize first if i wanted that? or can they coexist on metal? [19:11] ahh I see [19:11] ok, so to make best use of your resources [19:12] create lets say 6 kvm virtual machines on each server so you have a total of 12 vms between the two, make sure you provision them on a bridge to your local network so they are all on the same net between the two machines [19:13] make 1 of them your maas box, 1 of them your juju controller [19:13] that will leave you 10 vm boxes to check into maas and use as the machines you deploy your openstack bits to [19:18] i kinda suspected i'd need to run off of vms, not a big deal i suppose. Just wasn't a fan of virtualization on top of virtualization (since thats what im doing a bit now) [19:21] right but, for what you are trying to do with what you have .... [19:22] you could do it using lxd the same as vms [19:22] this might take a bit of lxd foo though, so be prepared [19:22] yea im reading up on lxd now (is it similar to how docker works in abstracting proesses?) [19:23] i've only used esxi/xenserver/kvm thus far [19:23] errr ... many would frown on that statement .... aside from googling, there are some good youtube vids on intro to lxd containers [19:24] lxd = system container, docker = process container [19:24] i guess i meant more in the "conceptual" idea of being like docker [19:24] they are both types of containers, yes [19:24] docker, more about the workflow [19:25] cool, well im gonna read up on lxd a bit then get provisioning [19:25] cool! good luck! [21:33] Bug #1787032 opened: [2.5] dhcpd.conf domain-name-servers has incorrect ip