[00:18] Bug #1669183 opened: [2.1.3] maas createds DNSResource records for discovered IPs, but doesn't appear to release them [00:31] how do i manually specify a pxe seed on a kvm guest to point at maas? I don't have maas managing my dhcp so i need to manually point the guest [02:27] Bug #1669213 opened: [2.2, trunk] Unable to change device IP because 'subnet' is listed incorrectly on 'edit'. [02:43] cholcombe I don't have access to my maas environment right now but once maas is installed you could always look at the dhcpd config to get what you need [02:45] Bug #1669217 opened: [2.2, trunk] Device Details doesn't surface error when updating interfaces [02:45] Bug #1669221 opened: [2.2, 5770] No block devices after commissioning [02:45] Bug #1669223 opened: [2.2, r5770] Commissioning Results has 'Status' column [02:54] Bug #1669217 changed: [2.2, trunk] Device Details doesn't surface error when updating interfaces [02:54] Bug #1669221 changed: [2.2, 5770] No block devices after commissioning [02:54] Bug #1669223 changed: [2.2, r5770] Commissioning Results has 'Status' column [02:57] Bug # opened: 1669217, 1669221, 1669223, 1669225, 1669226 [03:03] Bug #1669225 changed: [2.2, r5770, UI] Test Results outputs both stdout & stderr instead of a combined stdout/stderr [03:03] Bug #1669226 changed: [2.2, r5770] Commissioning 5 machines at the same time, 1 was set to Ready and the 'owner' wasn't cleared [03:15] Bug #1669225 opened: [2.2, r5770, UI] Test Results outputs both stdout & stderr instead of a combined stdout/stderr [03:15] Bug #1669226 opened: [2.2, r5770] Commissioning 5 machines at the same time, 1 was set to Ready and the 'owner' wasn't cleared [04:15] Bug #1669246 opened: [2.1, r5773] Custom commissioning script gets stuck in running state [04:15] Bug #1669248 opened: [2.2, 5773] Test are listed by script name, not user friendly at all === frankban|afk is now known as frankban [08:34] hello everyone, having an issue with Maas. The tft times out when I try to pxe a server [08:34] tftp* [09:01] is it only the tftp part that doesn't work? i.e. does the node get a dhcp offer? [09:15] yea dhcp is successful [09:15] it goes to the part that says tftp:///pxelinux.0 [09:15] then times out [09:16] I've tried to tftp to the Maas server from a separate server using standard tftp command line and same issue with timeout [09:19] HI All [09:19] could someone quickly point me to how I install python as part of the Ubuntu deployment? [09:43] jlec___: We add a line to late commands in the curtin_userdata file found on your MAAS controller at /etc/maas/preseeds/curtin_userdata [09:43] jlec___: http://pastebin.com/JvfxLDw3 [09:44] can I pull out IPMI addresses in bulk from MAAS using cli? [09:45] maas [env] nodes read don't seem to have it [09:52] zeestrat: thanks for this. Could you paste this somewhere else? Our corporate FW blocks pastebin. Even a query works for me [09:58] kklimonda^: I think I dumped all system_id's and then looped through with "maas [profile] node power-parameters [system_id]" or something like that. [12:52] Bug #1667574 changed: default gateway not set when deploying a machine with two interfaces in same network [13:13] Bug #1669425 opened: [2.2, rev5774] Error updating block devices after upgrade [13:13] Bug #1669428 opened: [2.2, rev5774] Invalid lshw data after upgrade [14:15] roaksoax, am I correct in remembering that you can set kernel parameters by using a tag? Can you point me to docs for that so I can try it out? [14:22] bladernr, you can change it globally in config AIUI [14:23] https://docs.ubuntu.com/maas/2.1/en/installconfig-kernel [14:23] https://docs.ubuntu.com/maas/2.1/en/manage-cli-advanced#specify-kernel-boot-options-for-a-machine [14:24] you supply kernel_opts to the tag oddly enough [14:24] yup, that's been possible since... 1.4? 1.3 ? [14:34] kiko, thanks! I remember roaksoax had mentioned it before, but didn't remember if I ever saw how to do it. much appreciated :) [15:16] Bug #1669460 opened: [2.2] Testing scripts don't have a name and the UI shows the scriptname [15:31] Bug #1669460 changed: [2.2] Testing scripts don't have a name and the UI shows the scriptname [15:37] Bug #1669460 opened: [2.2] Testing scripts don't have a name and the UI shows the scriptname [15:46] Bug #1669460 changed: [2.2] Testing scripts don't have a name and the UI shows the scriptname [16:35] anybody has examples of dhcp snippet? My interest is to have different IP ranges and gateway depending on the DHCP relay IP address that forwarded the request. [18:14] mimizone: you can use MAAS DHCP relay support in 2.2 :) [18:15] but 2.2 is just beta for now [18:15] mimizone: dhcp snippets is just any dhcp config you want to put in there [18:15] additional to what maas already generates [18:15] ahah having so much "fun" already with MAAS and Juju 2.1 than 2.2 sounds a bit scary :) [18:16] in 2.2, Maas is the relay itself? or it has the built in support for understanding the request comes from a relay? [18:20] kiko so I had the install fail again last night, and I grabbed logs for both a successful and failed install [18:21] stormmore, that's great! super plus great in fact! [18:21] stormmore, can we get a bug report filed with a precise description? [18:22] launchpad? [18:25] mimizone, https://docs.ubuntu.com/maas/devel/en/installconfig-network-dhcp#dhcp-relay [18:29] mimizone: no, we use a relay on the network [18:29] mimizone: so we basically send the config to the dhcp relay [18:32] ok I see. I was a bit confused by the doc, it didn't look different from the dhcp provider configuration I use today. [18:33] roaksoax: maas assume the dhcp relay is configurable then. what's the requirement? [18:33] roaksoax: today I use isc-dhcp-relay in cumulus networks [18:34] roaksoax: on the doc screenshot, I don't see how MAAS knows about the relay itself, it shows only the fabric vlan. [18:37] Bug #1669547 opened: [2.2, r5770] Missing cloud-init/curtin event log messages (errors during deployment could cause machine not to be marked failed deployment) [18:43] kiko, filed === frankban is now known as frankban|afk [18:49] Bug #1669548 opened: rack controller drops initial registration with region [18:51] thanks stormmore [18:51] stormmore, roaksoax has sent some questions [18:52] looking at the update [18:54] roaksoax, not sure what you mean "username" [18:55] stormmore, what version of MAAS incidentally? [18:55] 2.1.3 [18:57] stormmore, be aware that 2.1.4 is just out, though it doesn't seem to fix your issue -- I'm intrigued by the traceback saying that /etc/maas is not writeable [18:57] roaksoax, explain that code to me -- does rackd actually write to rackd.conf? [18:58] stormmore, well, actually, 2.1.4 has a fix for bug 1626654 [18:58] mup, c'mon [18:58] kiko: I apologize, but I'm pretty strict about only responding to known commands. [18:58] mup, bug 1626654 [18:58] kiko: Bug #1626654: [2.0] TFTP issues when rackd — Missing connections to 1 region controller(s). [18:58] I am still curious as to why it is a consistency failure too [18:59] I agree [19:03] did 2.1.4 hit stable? [19:05] Bug #1669548 changed: rack controller drops initial registration with region [19:08] Bug #1669548 opened: rack controller drops initial registration with region [19:08] Bug #1669555 opened: [2.2, trunk] Pods currently dont auto refresh [19:09] yes just [19:11] OK I know I will be installing at some point soon then :P [19:11] fyi I have been doing about 5 - 10+ maas installs / day [19:20] stormmore, what are you using it for? [19:39] kiko, I am building a bootstrap node that I can use for bootstrapping a data center [19:40] kiko, the plan would be to run MaaS inside a k8s cluster but need to bootstrap the k8s cluster first somehow [19:40] stormmore, yeah, the helm guys are working on a MAAS chart for the former, I assume you've seen that? [19:41] Bug #1669567 opened: [2.2, r5774] Machines don't transition to deploy (but they turn on) [19:41] Bug #1669568 opened: [2.2, r5774] Machines don't transition to deploy (but they turn on) [19:42] kiko, yeah I was looking at that plus I saw a couple of other potential implementation options out there [19:43] kiko: stormmore - just FYI - the charts exist today https://github.com/att-comdev/openstack-helm/tree/master/maas [19:43] i haven't used this personally, but when i was at the openstack ptg, i was looking at a console with a near-functional maas in k8s, and another console with a functional maas in k8s [19:44] 2 separate dc's [19:44] lazyPower, hehe yeah I was thinking about the att-comdev repos, they have another one that runs it in kubernetes without helm [19:45] stormmore: tbh it should be simple enough to just extract what you need from those gotemplates. [19:45] replace where applicable, and kubectl apply + profit [19:45] exactly, although I kinda like the idea of helm for dev tools [19:46] that said helm gave me hell the last time I tried to use it [19:47] i'm 50/50 on helm and how useful it is today. Its a good start, but needs more love imho [19:47] that plus the workloads need more vetting, i've encountered some broken charts [19:48] yeah that is basically where I stand on it, but the question then becomes do I want to roll my own or work with the helm team [19:48] six of one half dozen of the other [19:49] at the end of the day you're left to support the workload yourself. So choose which path gives you less heartburn [19:50] that is a very ops outlook :P I am thinking that working with a team that has the right idea but needs encouragement and help might be a better option ;-) [19:51] heck I probably would have given up on maas and juju somewhere in v1 if I didn't think that ;-) [19:51] no kidding. juju 1.1x was rough [19:51] I have a 1.21 cluster I have never touched because of the upgrade pain! [19:51] can I say understatement of the day :P [19:52] but then again 2.x is a light year and a half ahead [19:52] kiko, yeah I think that is a new build migrate path for least pain [19:53] ok first maas install of the day [19:55] stormmore, I'm going to get allenap to look at your bug tomorrow, he owns the registration code and will answer my question above [19:56] no problem, it is just an annoyance that a reinstall fixes right now [20:08] Bug #1669570 opened: 2.20-beta3+bzr5774: failed migration, cannot ALTER TABLE maasserver_node because it has pending trigger events

[20:10] probably going to upset my laptop with the size of VMs I am about to start :-/ [20:11] heh [20:12] nope a failed maas install saved me :-/ [20:13] I am getting another intermittent error after creating a superuser and then trying to log in [20:13] "The MAAS server gave an unexpected response: 503" has happened twice now [20:17] Bug # changed: 1654412, 1654432, 1654450, 1655721 [20:49] I put a small minute pause into my automation and that seems to have solved that problem and so far no sign of the rackd problem [20:51] I need to put a delay in anyway until the maas server is done importing the cloud img [21:14] where did they move the power templates to? [21:21] oh crap you guys removed that functionality :-/ [22:14] Bug #1669591 opened: [SRU] MAAS 1.9.5 [22:14] OK I am beginning to think MaaS 2.1 is going to kill my use of VBox and move to kvm / libvert [22:14] libvirt [22:37] that and the fact that trying to enlist 3 virtual machines at one time causes virtualbox to go into a weird locked up state [22:38] it appears that only one tag can provide kernel_opts. if a node has multiple tags, and each tag has kernel_opts values, only one is actually used at boot. is that something other experience? I use MAAS 2.1 [22:39] to make it work I had to put all the options I need at boot in one single tag as such "kernel_opts": "console=tty0 console=ttyS4 mei-me.disable_msi=1" [22:41] mimizone, pretty sure that's normal - https://docs.ubuntu.com/maas/2.1/en/manage-cli-advanced#specify-kernel-boot-options-for-a-machine [22:44] Bug #1602482 changed: [2.0rc2] Incorrect DNS records [22:44] Bug #1630679 changed: [2.1b1] Can't use custom image repository [22:44] Bug #1636861 changed: [2.1.1, trunk, bzr5510] UI error when adding a chassis [22:45] pmatulis: can beat the documentation to clarify that :) [22:45] can't