[05:53] blake_r: thanks, will do [05:53] roaksoax: thanks for the help [07:34] blake_r roaksoax: FYI, bug is at https://bugs.launchpad.net/maas/+bug/1563701 [07:38] Bug #1563701 opened: [2.0] VLAN interfaces of secondary rack-controller are not reported [07:47] Bug #1563701 changed: [2.0] VLAN interfaces of secondary rack-controller are not reported [07:53] Bug #1563701 opened: [2.0] VLAN interfaces of secondary rack-controller are not reported [09:15] g'day peeps, just thought I'd drop in and ask a couple of questions if no one minds. Firstly, I do understand virtualisation and running OSs in a VM, although I only used VMWare many many years ago. [09:15] What I really want to know, is MAAS designed purely for cloud services or any server services? Can I use it for generic LAMP stuff, DNS, DHCP servers and the like? [10:38] Bug #1563779 opened: [2.0a4] maas-rackd missing presumed lost [10:44] Bug #1563779 changed: [2.0a4] maas-rackd missing presumed lost [10:47] Bug #1563779 opened: [2.0a4] maas-rackd missing presumed lost [11:38] hellow [11:38] Bug #1563799 opened: [2.0a4] Permission error for boot-resources/cache [11:38] Bug #1563807 opened: Systemd units need to reflect updated MAAS names [13:27] Bug #1563859 opened: [2.0a4] Installing maas-region-api fails with missing maascli module [14:29] hi there, first time around [14:30] guys i got a seriuos hardware failure in my all-in-one maas server... is there any way to move the active nodes to the new server? [14:36] I have an small maas server with 2 NICs, one with direct link to internet and the other connected to a 24 port switch where all the server are linked. With yhis maas server down i think no one server will boot again :( [14:37] I will setup another maas server, but i dont know if there is a way to reconnect the servers with the new maas controller [14:56] Is there any documentation on issues with bootstrapping a node in juju? I'm trying to find the location of some sort of logfile that can give me details as to why juju isn't bootstrapping an environment on a particular node [14:57] I'm using maas and have 4 nodes in Ready state, but I can't seem to get one of the nodes to be bootstrapped with juju [16:41] Hello all. I'm in need of some help with maas [16:42] I have setup the cluster to manage DHCP and DNS on eth0 [16:43] I'm at the stage were I pxe boot a node [16:43] It reaches the maas-enlisting-node login: [16:44] And nothing happens after that. It should right some more stuff on the screen and then shut down. But it doesn't [16:44] It shoul write not right [16:51] hmm [16:51] you need a backdoored image to figure out what's wrong [16:51] ikat, https://maas.ubuntu.com/docs/troubleshooting.html#debugging-ephemeral-image [16:52] bugrum, if you do bootstrap --debug you get pretty copious output [16:52] mantrado, you can absolutely use MAAS for anything which runs on a bare-metal OS [17:05] kiko, Thanks but I'm not sure how to use that info [17:05] ikat, you need to backdoor the images and then run enlistment again [17:05] ikat, when you do you will have a login to enter the machine [17:05] which will let you inspect to see what's wrong [17:05] I'm fine with the first command to install the software but what do i do with the rest of the lines [17:07] kiko, What do I do with this? It looks like code I need to put somewhere [17:07] imgs=$(echo /var/lib/maas/boot-resources/*/*/*/*/*/*/root-image) for img in $imgs; do [ -f "$img.dist" ] || sudo cp -a --sparse=always $img $img.dist done [17:09] ikat, you need to run this on your maas server, which hosts the images [17:09] but I'm assuming you understand how MAAS works :) [17:09] Where does MAAS store the images? In postgresql? [17:10] so I have been using MAAS for a few months now, I keep running into network setup problems on my nodes due to the fact that I have multiple gateways [17:13] maaku, yes on the region, and then it's written to the filesystem on the cluster [17:14] stormmore, I've heard about this problem before, how could we make it better? [17:14] kiko: is it wise to put 0.5GB binary blobs in sql? [17:14] anyway just seems weird. thanks [17:15] kiko I am thinking that when you assign a gateway to a managed network that it should when it deploys the node setup a new routing table with a default route to the gateway [17:15] maaku, using the blob store? it works acceptably, and we get HA for free out of it [17:15] stormmore, yes, basically you want the ability to specify one or more static routes per subnet I think? [17:15] kiko: I did it manually on one node so I can access a web service on the right IP [17:16] right [17:16] stormmore, it's on the roadmap for post-16.04 [17:17] kiko: well at least a default gateway per subnet, even though I fixed the one node when I went to deploy more it was still having issues accessing the first node (I haven’t troubleshoot that yet, plan on manually deploying the nodes and hand building the network config to determine where that failure is) [17:17] stormmore, hmm, mpontillo, blake_r: does 2.0 change the default gateway setting? [17:17] I don't get it. [17:18] kiko: static routes aren’t sufficient cause the web service might be accessed from an unknown range [17:18] ikat: what don’t you get? [17:18] This is a command "sudo apt-get install --assume-yes bzr" [17:19] this is also a command "bzr branch lp:~maas-maintainers/maas/backdoor-image backdoor-image" [17:19] What is the rest "imgs=$(echo /var/lib/maas/boot-resources/*/*/*/*/*/*/root-image) for img in $imgs; do [ -f "$img.dist" ] || sudo cp -a --sparse=always $img $img.dist done for img in $imgs; do sudo ./backdoor-image/backdoor-image -v --user=backdoor --password-auth --password=ubuntu $img done" [17:20] another chain of commands [17:20] kiko: for the record, the web service I am starting with is juju-gui [17:22] stormmore, I'm not sure I understood your last point, but I do know that the networking arch in MAAS has changed with 2.0 [17:22] Cool. So maybe I got it to run. Now what? Just repeat the enlisting effort? [17:24] cool backdoor and ubuntu. I'm in [17:24] kiko, stormmore: I'm out today so don't have the code in front of me, but IIRC, no, we use the default gateway defined on the subnet [17:24] kiko: cool I will review the notes on 2.0 then and see if it is changing in a way that will resolve my issue. The long and the short of it is that the nodes need to be able to respond on multiple interfaces with different gateways regardless of what the client IP is. [17:25] mpontillo: it would be interesting to see how you are doing that, the 2 methods I know are either setup a routing table per subnet or use metrics on each gatway [17:27] I prefer the former as it just seems cleaner and more logical to me. in my environment I have 2 NICs, 1 which will be dedicated to all the management traffic and thus needs to have a gateway to get to the repos, etc. and the other will have all the public traffic which has a completely different gateway (right now the first gateway is the MAAS server) [17:28] stormmore: we don't do metrics or static routes yet, we select a single gateway =( would love your feedback on how it should work. I am aware it's an issue. [17:31] ikat, so try and figure out what isn't happening that should -- do the logs suggest a problem? [17:31] mpontillo: http://paste.ubuntu.com/15555049/ has the 2 files that I had to change to get my node to communicate on both default gateways [17:34] mpontillo: since I assume that MAAS is writing the /etc/network/interfaces file I would assume it should be able to add the appropriate post-up lines for the different subnets [17:34] I'm going through them now. It all looks quite normal. More on this tomorrow as it's time to go home from work. Thanks for the pointers kiko! [17:35] mpontillo: but it would also need to update /etc/iproute2/rt_tables to add the additional routing tables [17:39] ikat, thanks -- cool, let me know, and in particular, ping smoser if you have an update [17:40] kiko, will do [17:48] Bug #1563990 opened: Need a GUI option to set per node kernel parameters

[17:59] stormmore: thanks; I'll bring this up with the appropriate people [18:01] mpontillo: anytime, I am really enjoying the tools and want to make them better, MAAS has already caused us to hire a contractor for our lab to deploy it further than my little test net [18:06] So. I've set up a second network/zone/cluster, but I cannot seem to get IPMI or WoL working. (Both work on my primary zone controller.) IP addresses are allocated fine; IPMI and WoL both work from the secondary zone controller when executed manually. I tried sniffing, and don't see any packets going to the MAC or IP, at all (after the PXE boot stuff). Nor do I see any pertinent logfiles. Any ideas? [18:37] stormmore, that's very cool [18:38] Ravenpi, hmm, sounds like some misconfig [18:39] That's what I was thinking, but I've gone over what I can think of. And since PXE works 'n stuff, it's not broken-broken. [18:40] But I'm not sure where to go from here. Do you know which process it is that sends out the IPMI or WoL stuff? I'd be happy to strace it. [18:43] so [18:43] AIUI everything on the cluster controller is launched from maas-clusterd [18:43] PXE will work if DHCP and TFTP are running on the cluster [18:49] Thanks. I'll go a-digging. [22:03] Installing on 16.04 beta2, base install + maas -- "Internal server error" because `bzr` was not installed [22:04] I had to apt-get install bzr [23:10] Bug #1556137 opened: Juju 2.0: cannot connect to LXC container when deploying bundle with LXC container placement with maas 1.9.1 as provider <2.0-count>