[03:14] hey, is it possible to test MASS on a virtualbox ? [03:39] Bug #1560789 opened: [1.9.1]failed deployments because of accidental boot device selection possible and can't deselect === blahdeblah is now known as Killfirefox === Killfirefox is now known as blahdeblah === beisner- is now known as beisner [05:06] I am installing MAAS on ubuntu 14.04, but getting an error like this: [05:19] ImportError: Could not import settings 'maas.settings' (Is it on sys.path? Is there an import error in the settings file?): No module named maasserver.config dpkg: error processing package maas-region-controller-min (--configure): subprocess installed post-installation script returned error exit status 1 [05:20] dpkg: error processing package maas (--configure): dependency problems - leaving unconfigured Errors were encountered while processing: maas-region-controller-min maas-dns maas-region-controller maas E: Sub-process /usr/bin/dpkg returned an error code (1) [05:20] Please help with this issue [06:01] Bug #1560815 opened: [2.0a3] apt-get install maas fails due to postgresql connection [06:13] Bug #1560815 changed: [2.0a3] apt-get install maas fails due to postgresql connection [06:19] Bug #1560815 opened: [2.0a3] apt-get install maas fails due to postgresql connection [07:01] Bug #1560830 opened: maas power check won't work with ipmi hosts with the string 'on' in them [09:20] roaksoax: ping, are you around yet? [09:47] Hi! Is it possible to configure a node, so that I can manually startup a machine? [10:34] Bug #1560901 opened: install maas (2.0 alpha3) from image (ubuntu 16.04) does not include power-types [12:45] voidspace: here [12:53] Bug #1560901 changed: install maas (2.0 alpha3) from image (ubuntu 16.04) does not include power-types [13:01] roaksoax: morning [13:14] Bug #1560971 opened: after default installation of maas the rack and region controller are not connected [19:01] Hi, can MAAS deploy CentOS servers? and do I need DHCP for it to work? [19:02] Hi, can MAAS deploy CentOS servers? and do I need DHCP for it to work? [19:10] Hi, can MAAS deploy CentOS servers? and do I need DHCP for it to work? [19:10] Hi, can MAAS deploy CentOS servers? and do I need DHCP for it to work? === bogdanteleaga_ is now known as bogdanteleaga === mup_ is now known as mup [19:27] Hi, can MAAS deploy CentOS servers? and do I need DHCP for it to work? [19:27] Hi, can MAAS deploy CentOS servers? and do I need DHCP for it to work? [19:27] Hi, can MAAS deploy CentOS servers? and do I need DHCP for it to work? [19:30] yes and no [19:57] in MAAS 2.0, is there a method to reconfigure the DHCP scope for a vlan once it's enabled? The options disappear once you initially add them [19:58] I had pointed the gateway at the network gateway, but the nodes are attempting to query a url on that address during initial commissioning, so I am assuming I was supposed to point it at the MAAS server itself [20:26] I fixed it from the CLI - maas admin subnet update 10.101.0.0/16 gateway_ip=10.101.0.X, where X is not the network gateway but is the address of the PXE interface on the MAAS server [20:45] LiftedKilt: the gateway shouldn't really impact in any way where the nodes should be communicating the the rack/region controller [20:45] LiftedKilt: if that's the case, there might be a bug [20:46] roaksoax: It doesn't seem like it should, but it was attempting to query http:///meta-data/instance-id or something like that [20:47] LiftedKilt: make sure that /etc/maas/rackd.conf is pointing to a correct IP other than localhost [20:47] I changed the gateway to be the ip of the maas controller, and rebooted and the nodes started enlisting no problem [20:47] LiftedKilt: that may be the reason [20:47] LiftedKilt: strange indeed. That seems like a bug to me [20:49] LiftedKilt: when you say "DHCP scope" are you talking about the IP ranges, gateway IP, or both? both are intended to be editable on the subnet details page, but that functionality hasn't been finished yet for the 2.0 alpha release [20:49] LiftedKilt: the "Provide DHCP" option on the VLAN details page is intended mainly to specify the rack controllers, but we added initial IP range and gateway specification as a convenience feature as well [20:50] roaksoax: the rackd.conf file is pointing at the right ip [20:50] LiftedKilt: if those items already exist the 2nd time you "Provide DHCP", the mechanics of allowing the user to edit them in that tiny drop-down form are a bit problematic [20:51] (also, when the rack registers it attempts to discover the default gateway and assign it to the subnet, FYI.) [20:51] mpontillo: yeah that's the issue I was seeing [20:53] mpontillo, roaksoax: it's odd that changing the network gateway to the IP of the maas server was necessary for enlisting though [20:53] LiftedKilt: that shouldn't be necessary. The rack / region controller communication should determine what IP address to give the node for the metadata server [20:54] LiftedKilt: I'm a bit confused about that part. is your region controller on a different subnet; one that would require the gateway to access? (the metadata server is on the region controller...) [20:55] So I did a clean install of maas 2.0 alpha3, added the 16.04 image for commissioning/deployment, and tried to pxe boot a node [20:56] mpontillo: everything pxe is on a single subnet, I have a separate subnet for access to the maas controller [20:57] on maas 1.9 I had the webui listening on the 192.168.4.0/24 network, and the api access for pxe on the 10.101.0.0/16 network [20:58] LiftedKilt: so? regiond (ethX on 10.101.0.0/16) <---> (10.101.0.0/16 on ethX) cluster (ethY 192.168.4.0/24) --> nodes [20:59] LiftedKilt: and the machines would PXE boot off 192.168.* , and told that the API( metadata server) was on 10.101.* ? [21:00] LiftedKilt, do I understand correctly that there is no routing between the PXE subnet and the management network? [21:00] machines would pxe off of the 10 network, and would stay on the 10 network post deployment. The 192.168 network was simply for ssh access to the box and access to the webui [21:03] LiftedKilt: I just checked the code; it looks like we just pull maas_url out of /etc/maas/regiond.conf and tell newly enlisted nodes to use that to report back to MAAS [21:04] mpontillo: that's the case if we cannot detect from what rack controller we pxe booted from [21:04] that's get_preseed_context() in maasserver/preseed.py; 'metadata_enlist_url': absolute_reverse('enlist', base_url=base_url), --> absolute_reverse() just grabs the URL from the config and appends the desired enlistment URL [21:05] mpontillo: but if we do, we tell it the IP from which the rack controller was able to connect through [21:06] mpontillo: if rack_controller is None: [21:06] base_url = None [21:06] else: [21:06] base_url = rack_controller.url [21:06] mpontillo: is not maas_url straight out of regiond.conf [21:07] roaksoax: well, the mechanism looks slightly different depending on which URL you're talking about; I see that the node has a chance to override some properties, but it isn't clear to me which ones [21:08] roaksoax: but I see what you're talking about [21:12] mpontillo: yeah, so if we can't determine the url, because say rackd.conf points to localhost/MAAS , then we give maas_url of regiond.conf [21:38] good morning! [21:39] I'm getting erros during commissioning - is there any way to debug it or get the logs from the screen? [21:39] I guess the cloud-init stuff fails somewhere [21:41] ReSam: you may be able to see what's failing on the event log under the machine details [21:41] ReSam: if not, you'll have to connect to a serial console and see what's going on [21:41] roaksoax: nope - it just says "Failed commissioning" [21:41] I can see the screen of the machine via VNC - but the text scrolls past quickly [21:42] or is there any way to ssh into the commissioning ubuntu 14.04 [21:43] ReSam: you can, but only if you are using maas 1.9+ [21:43] yes - please tell me! [21:44] I can connect via SSH, but my key (which I entered in maas for my user) is not accepted [21:47] roaksoax: or is there a default password I don't know yet? [21:51] ReSam: on commissioning, there's an option there to import your ssh key [21:51] ReSam: so just ssh ubuntu@ [21:51] I imported my key - doesn't work. [21:51] I can ssh - but it rejects my key - and doesn't allow me to enter a password [21:52] ReSam: then it might be failing before getting to the point, because the option should prevent the machine from turning off after commissioning [21:53] yes - it stays on. I can see it via VNC from my management system [21:55] ReSam: so, you are clicking on " Allow SSH access and prevent machine from powering off" ? [21:55] roaksoax: yes [21:55] ReSam: strange... since after running the commissioning image it should have imported your key [21:56] ReSam: and prevented the machine ffrom powering off [21:56] but it doesn't accept my key. it just tries all 5 keys - and then says "Permission denied (publickey)." [21:56] ReSam: right, which means it may have not been imported [21:57] ReSam: if you go to the machine details -> Latest machine events -> view full history [21:57] ReSam: hwat does it show ? [21:58] roaksoax: Node changed status - From 'Commissioning' to 'Failed commissioning' [21:59] ReSam: how quickly ?\\ [21:59] what's the timestamp since it started ? [21:59] 3min or so. I can watch it doing stuff. but it's to fast to actually read stuff [22:00] It looks like apt-get install fails or so... [22:00] ReSam: if apt-get fails, are you runing a proxy ? [22:00] ReSam: is maas-proxy running ? [22:00] yes - should all be working [22:02] roaksoax: I was hoping I could find logs in: /var/log/maas/rsyslog [22:02] but it is empty... [22:02] ReSam: no, cloud-init is not configured, there's an open bug for that [22:03] mhm - but it prints some message that it passes "log_host=" as parameter during commission... [22:04] ReSam: yeah, but cloud-init doesn't take the kernel parameters for that [22:10] Bug #1561222 opened: [UI 2.0a3] Event's table spacing is incorrect [22:20] ReSam: in : /etc/maas/templates/commissioning-user-data/user_data_config.template do: [22:21] ReSam: add at the bottom: http://pastebin.ubuntu.com/15483272/ [22:21] ReSam: that should give you rsyslog [22:21] roaksoax: thanks, I will try it. [23:03] how do you delete ip ranges in maas 2.0? [23:04] I can create reserved and dynamic ranges, and I can read them, but there's no cli option for deletion [23:07] LiftedKilt: maas iprange delete [23:07] mpontillo: I'm just seeing that now - why the difference between the iprange and the ipranges commands? [23:07] LiftedKilt: it's consistent with the other APIs; there is a plural endpoint for creation and listing, and a singular endpoint for doing something with an individual object [23:08] mpontillo: ok cool. Good to know - thanks! [23:08] LiftedKilt: do "maas --help" to see what I mean - there are basically two endpoints for each object [23:08] np [23:12] ReSam: did that work ? [23:13] roaksoax: sry - other things need to be fixed first... I will report back tomorrow. thx.