[01:08] Bug #1437092 was opened: Username not logged in machine de-allocations [01:08] Bug #1437094 was opened: Sorting by mac address on webui causes internal server error [01:31] Howdy! [01:33] I posted earlier in the day about having trouble with grubx64.efi. I have narrowed down the problem to a specific adapter "82599ES 10-Gigabit SFI/SFP+" that the grub image is not able to support. [01:34] I can commission machines if I UEFI off of the integrated adapter. [01:37] My question is how does maas build grubx64.efi (with grub-mkimage and what options)? [02:04] anyone on? [07:22] hey all. does anyone have any experience getting MAAS working in Virtual Box instances? I'm having some problems with the PXE stuff and would like some pointers. if you can assist then please pm me as the noise in the channel is likely to mean I miss your post. === racedo_ is now known as racedo [09:25] Bug #1437221 was opened: regiond.log contains log entries that belong to maas.log [09:55] Bug #1360306 changed: Feedback of bulk actions is lacking [09:55] Bug #1437231 was opened: No instant feedback when performing an action in the UI [10:55] Bug #1437251 was opened: "ImportError: cannot import name node_query" when running test.region with some test modules, but not when passing a folder or running without arguments [12:16] Bug #1437288 was opened: 1.8 intermittently fails to capture lshw output when commissioning [12:28] Bug #1437288 changed: 1.8 intermittently fails to capture lshw output when commissioning [12:34] Bug #1437288 was opened: 1.8 intermittently fails to capture lshw output when commissioning [12:46] Bug #1437288 changed: 1.8 intermittently fails to capture lshw output when commissioning [13:19] Howdy! does anyone on know how MAAS downoads, modifies, and/or creates grubx64.efi ? [13:21] kiko, do you have a minute for a question on MaaS 1.7 and DNS? [14:22] Bug #1437280 was opened: maas 1.8.0alpha7 two machines have same ip address

[14:22] Bug #1437353 was opened: UEFI network boot hangs at grub for adapter 82599ES 10-Gigabit SFI/SFP+ [14:52] Bug #1437364 was opened: node that was never previously commissioned in ready state [14:55] nodtkn: greubx64.efi is obtained from the ubuntu archives [14:55] nodtkn: can you give me the link of your bug again please? [14:56] nodtkn: also, that could potentially be an issue with grub, rather than maas [15:16] Bug #1437364 changed: node that was never previously commissioned in ready state [15:40] Bug #1437380 was opened: Form validatation on add node inconsistant [15:40] Bug #1437383 was opened: button text on add node is inconsistent and out of date [15:40] Bug #1437388 was opened: exceptions.AttributeError: 'NoneType' object has no attribute 'is_superuser' [15:55] Anyone have an idea as to why my nodes are not showing up in DNS? [15:55] I am running MaaS 1.7 stable from the PPA === rvba` is now known as rvba [16:16] mrhatch, ah, you're back [16:17] mrhatch, just to be sure, are you using DNS managed by MAAS? i.e. in your cluster interface, do you have both DHCP and DNS set to be managed? [16:17] yes I do [16:18] mrhatch, okay. and are you sure you are sending your DNS queries to your maas server? [16:18] i.e. what does nslookup tell you your default resolver is? === chuck_ is now known as zul [16:19] I think I see the problem [16:19] but I don't know how to fix it [16:20] The server is set to my Router address [16:28] Should that be set on the networks tab? [16:31] roaksoax: I have opened bug https://launchpad.net/bugs/1437353 my guess is that this is either a grub or bios issue but could also be related to mishandling of tsize by the tftp server [16:34] roaksoax: my plan is to gather tcpdumps and log files for two different machine manufactures with the same network adapter [16:35] mrhatch, the resolver is configured in your local linux environment [16:35] mrhatch, what machine are you issuing the queries on? [16:35] mrhatch, are you on the maas server itself? [16:35] if so [16:38] right now I am on the maas server itself [16:39] is the dns setup done through /etc/network/interfaces or via networkmanager? [16:39] I just did a plain install here is what I did: http://brd-dumb.com/?p=17 [16:40] mrhatch, did you install Ubuntu server or desktop on the node? [16:40] server [16:41] looks like /etc/network/interfaces has the only dns-nameserver pointed to the router [16:41] which explains the problem [16:41] right [16:41] change that to be 127.0.0.1 [16:43] should I add anything for resolving outside of the network? [16:43] like google? [16:44] mrhatch, if you configured your maas upstream dns server correctly, 127.0.0.1 should just work [16:44] in the maas options you can specify your upstream dns server (which could be your router's) [16:45] (normally even without the upstream dns server it should work) [16:46] upstream is working fine, still can't resolve the node [16:46] it is coming up with localhost as the resolver [16:46] would I need to reboot or reload the node at this point?] [16:49] That seemed to take care of it...the reboot that is [16:51] kiki, rebooting the node seems to have fixed things, I can now do the lookups from the host [16:54] mrhatch, okay, cool -- the reboot probably was avoidable, but I'm happy it's sorted now :) [16:54] I may have one other question...I have noticed issues with 1.7 and virtualbox machines last night [16:54] I am hoping that getting the dns sorted out may have fixed it [16:55] mrhatch, okay, keep us posted [17:15] Ok, I have tried enlisting two different virtual box nodes and they go through all of the steps to commission but never seem to finish [17:16] what do their consoles show you? [17:16] anything suggestive [17:16] The node list says [17:17] failed deployment [17:17] I can't get them to boot up [17:17] they keep going back out to the pxe server [17:17] mrhatch, the commissioning will make the machines boot up -- are they not booting up at all? [17:19] they boot up, but try to load the OS and then shutdown [17:20] mrhatch, what happens when they try and load the OS? [17:20] there's something weird there [17:23] basically they are stuck in an endless reboot loop [17:23] Bug #1437424 was opened: Main navigation and action bar should be sticky [17:23] Bug #1437426 was opened: No view for loading page or notification for connection error [17:23] I get to through commissioning, to acquire and start [17:24] they seem to do everything I expect a node to do but never complete the reboot [17:24] mrhatch, oh, the machines actually commission correctly? [17:24] and you can tell them to start? [17:26] yes I tell them to start, they boot up as normal, I get the login prompt and then faster than I can read they shutdown [17:26] mrhatch, when you "start" node or when you "commission" node [17:27] They seemed to work with 1.5 just fine [17:28] when I start the node...virtual box doesn't seem to actually turn a node on when you interact with the maas interface [17:41] Ihave been powering them on and off manually [18:14] Bug #1437443 was opened: MAAS needs to storage failed commissioning or successful commissioning log just like we do witht he isntall log. [18:23] I was able to watch the virtualbox machine a little closer, it looks like it did everything it was supposed to on start, but that final reboot after the completion for some reason never triggers the status to change in MaaS correctly [18:35] hmmm [18:35] that's odd === mup_ is now known as mup [19:12] so I looked in the install log on MaaS and found Installation failed with exception: Unexpected error while running command. Command: ['curtin', 'curthooks'] Exit code: 3 [19:14] hmm [19:21] I am going though the different network cards the vm can emulate right now to see if that makes a difference [19:24] it's definitely odd [19:24] any reason why you want to use virtualbox and not virsh? [19:25] I am just testing at home and only have limited hardware...I don't want to wipe out my windows box to have machines to play with juju [19:25] ah [19:25] interesting use case ;) [19:25] well I work for a small company...they want some big things but don't want to spend big $$ [19:26] before I can go to the boss with this is great, I need to work out all of the kinks [19:26] and have something to demo [19:27] right [19:27] I was mostly noting that instead of using virtualbox you could use virsh [19:27] which should be equivalent [19:27] and is reported to work with maas [19:27] roaksoax, ^^ [19:29] virtualbox worked on 1.5 [19:29] I wonder if it worked with d-i [19:29] mrhatch, you can try one thing [19:30] mrhatch, click on the "use debian-installer" button [19:30] to use the same provisioning mechanism as the default in 1.5 [19:30] I will spin up another vm and give it a try...I also just started using the 32 bit version of Ubuntu for testing purposes [19:31] mrhatch: your log doesn't give any idea as to what it might be failing. You should be able to see the install log on the web ui [19:31] the full one [19:32] mrhatch: that wbeing said, we dont officially support VBox either [19:32] so might be related to the configuration of your VM [19:32] mrhatch: also, what version of MAAS and curtin convinationsa re bieng used [19:32] give me a few and I will check [19:34] maas is version1.7.2 [19:35] mrhatch: what about curtin version in the maas server 1.X bzr 201? [19:37] where would I find that? [19:37] mrhatch: in the maas server, dpkg -l | grep curtin [19:38] ii curtin-common 0.1.0~bzr201-0ubuntu1~14.04.1~ppa0 all Library and tools for curtin installer ii python-curtin 0.1.0~bzr201-0ubuntu1~14.04.1~ppa0 all Library and tools for curtin installer [19:46] the only real difference that changing it to using the debian installer is that it no longer marks itself as a failure, but keeps the status of deploying [19:47] mrhatch: maybe it is still deploying [19:47] mrhatch: twhat's the full curtin log? [19:47] I just wiped out those nodes, give me a few and I will get you logs from the next deployment. [19:59] http://pastebin.com/nQELWT8G [20:00] http://pastebin.com/bqn5f7SS [20:00] roaksoax do those logs help any? [20:01] mrhatch: yup [20:01] seems lie a bug in curtin [20:01] not MAAS [20:01] mrhatch: can youplease file a bug? [20:04] I will gladly file the bug if you point me in the right direction [20:04] I am guessing that the version of curtin or the use of curtin has changed since MaaS 1.5? [20:05] mrhatch: launchpad.net/curtin [20:05] mrhatch: yup [20:06] is there anything other than it is failing with Virtualbox I should include? [20:08] mrhatch: all the information of what type of disk it is being virtualized and stuff like that [20:09] if you think it happens only in 201 (and not in 196), can you roll back to 196 and verify that ? [20:09] it *really* seems not likely to me that it regressed there. [20:10] OK [20:12] 196 being 1.5 of Maas? [20:12] Oh you mean curtin-common [20:12] right. [20:13] if this is very reproducible for you, it'd be nice if we could get an 'lsof' right after it fails [20:15] crap, I didn't realize that taking out curtin-common and its dependencies would take out maas as well [20:29] smoser: since the vm is not up long enough to even log in I don't think I can do that [20:30] mrhatch, oh no. [20:30] sorry. you can / cloud hav just installed (dpkg -i) the curtin_common and python-curtin) [20:31] those are likely the 2 you have installed [20:31] you can ake the vm mnot reboot [20:31] the funny thing is the machine comes up and shows a login screen but that only lasts for about 20 seconds [20:31] /etc/maas/preseeds/curtin_userdata [20:32] that file has [20:32] power_state: [20:32] mode: reboot [20:32] if you comment it out, you'll just not reboot. [20:32] you should be able to ssh into the install environment too. [20:32] thats possibly bug, but its true :) [20:33] I'll give that a try, but I have to rebuild my maas server since I yanked curtin out and it uninstalled some of maas [20:56] smoser I managed to get into the vm that was failing, is there anything you would like to know? I had to boot to root mode and set a username and password to get in [20:57] well, during the install, you should be able to ssh in as the user that did the deploy [20:57] with their keys [20:57] err.. their keys, ubuntu user. [20:57] anyway... youf igured it out. [20:57] i assume that the directory is unmounted now ? [20:57] mrhatch, can you [20:57] apt-get install pastebinit [20:57] yeah that was one of the things, the user was never created [20:58] (and do you avhe internet or proxied internet from there ?) [20:58] sure I can do that [20:58] the uubntu user shoudl be ther.e [20:58] straight internet [20:58] ok. [20:58] It wasn't there...I tried to just set the password and it wasn't there so I created it, set a static ip and then rebooted and I am in as ubuntu now [20:58] so pastebinit /proc/mounts [20:58] mrhatch, ah. right. the user wouldnt be crated at that point. [20:59] but if you cause the install to not reboot, there is an 'ubuntu' user there tha tyou could go in as [20:59] yeah... this wont be terribly helpful [20:59] so what happened... [20:59] er.. heres what happend. [20:59] curtin does install [20:59] then chroots and runs 'update-grub' [20:59] and then [20:59] unmount /target [21:00] which is where it fails [21:00] but in your case, something got left behind during hte chroot [21:00] some filehandles were left open [21:00] and curtin's unmount failed [21:00] i'm not sure what gets done, keeping that filesystem busy. [21:00] :-( [21:07] ugh evidently ssh server is not installed no good way to get data out of the vm... [21:08] wife is ready for dinner, I am gonna head out thank you all for your help. I will try this again after my maas server is rebuilt [21:20] Bug #1437491 was opened: Cannot poweroff node when they are in New state but ON [21:20] Bug #1437494 was opened: Cannot view power details from CLI for New nodes as Admin user. But I can view them in the WebUI