=== CyberJacob is now known as zz_CyberJacob === zz_CyberJacob is now known as CyberJacob === CyberJacob is now known as zz_CyberJacob [02:07] Bug #1441933 was opened: Internal Server Error when saving a cluster without Router IP [06:26] rvba, ping === zz_CyberJacob is now known as CyberJacob === CyberJacob is now known as zz_CyberJacob [08:06] dimitern: hey [08:07] rvba, hey, it's fine - I needed to confirm that missing lshw data always leads to failed deployments, but roaksoax already did this [08:08] dimitern: okay, cool :). [08:08] rvba, which means juju should indeed fail when lshw is missing/unparsable [08:09] Right. [10:21] Bug #1442059 was opened: 1.8b1 Failed deployment timeout powering on AMT [10:33] Bug #1442059 changed: 1.8b1 Failed deployment timeout powering on AMT [10:39] Bug #1442059 was opened: 1.8b1 Failed deployment timeout powering on AMT === wojdev_ is now known as wojdev [12:09] Bug #1442112 was opened: booting ppc64el with hwe-u kernel hangs in initrd [12:39] Bug #1442131 was opened: 1.8b1 redundant owner column on devices page === zz_CyberJacob is now known as CyberJacob [13:51] Bug #1442157 was opened: Manually added node doesn't get added to MAAS managed DNS [14:21] Bug #1442162 was opened: Spurious test failure: maasserver.api.tests.test_nodes.TestFilteredNodesListFromRequest.test_node_list_with_ids_orders_by_id [15:46] Bug #1442234 was opened: maas-regiond and maas-clusterd clobber logs each time they're started [17:01] Bug #1442280 was opened: The ordering of the statuses changes when nodes change status [17:04] Bug #1442280 changed: The ordering of the statuses changes when nodes change status [17:13] Bug #1442280 was opened: The ordering of the statuses changes when nodes change status [17:16] Does anyone know if the 1.7.1 deb files for trusty around somewhere? the stable ppa has 1.7.2 which is not working for me, but 1.7.1 is in another install [17:41] serverascode: why is it not working? [17:41] serverascode: what are you affected by? [17:42] seems like the region controller doesn't auto configure, " Failed to connect to region controller, cannot upload leases", I don't see interfaces in the cluster in the gui [17:43] to be honest I don't really have time to troubleshoot it, I have a feeling 1.7.1 will work [17:44] I'd really like to use and recommend maas, but so far it has been quite the struggle [18:05] Morning folks. It seems like MaaS by default installs on the first disk. "d-i partman/early_command string debconf-set partman-auto/disk `list-devices disk | head -n1`" -- Would I just edit this to something more like "d-i partman-auto/disk string /dev/sdb" to specify the disk I want? I am using Curtin, rather than debian installer. [18:06] bmorriso: if you using curtin then you will want to change the curtin_userdata preseed [18:07] I don't see anything in there specific to partitioning/disk layout [18:09] bmorriso: that is because by default curitn picks the first disk [18:09] bmorriso: http://paste.ubuntu.com/10784573/ [18:09] bmorriso: add that to the config [18:09] bmorriso: change the /dev/sdb to the device you want [18:10] Is there a certain place that should go? [18:10] bmorriso: it can just go at the end [18:11] serverascode: did you ever get the cluster connected to the region [18:11] Thank you very much!! [18:11] bmorriso: no problem [18:16] blake_r: if I wanted to target a specific host with that command? According to this https://maas.ubuntu.com/docs1.7/development/preseeds.html I would do something like "curtin_userdata_ubuntu_amd64_generic_bfs2.poc" ? [18:16] With Curtin is there a way to target system information? Either brand of server, or type of disk? [18:17] I can give more context if it'd help [18:17] curtin_userdata_ubuntu_amd64_generic_trusty_bfs2 [18:17] don't include the fqdn [18:17] just the hostname [18:17] bmorriso: no real way of doing that [18:17] bmorriso: you could tag the machines in maas [18:18] bmorriso: then using the templating system we use, make a conditional on that tag [18:18] So tag it UCS/LUN or something similar, then based on those tags, match on the conditional with an "if/else" and I could set the partition that way? [18:19] I really don't want to get into managing preseeds per server, rather I'd like to say "If this has a NetApp disk (LUN) or If this is UCS booting from SAN, install Ubuntu on this disk, else install on the first disk" [18:36] bmorriso: no not that way [18:37] bmorriso: if you need to do something like that, then you will need to add that logic in the preseed to be processed by the client [18:37] bmorriso: I was speaking of tags in maas, like tagging a machine [18:39] I'll have to look at the docs to see. [18:41] blake_r: I used what you suggested for /dev/sdb but it actually installed it on /dev/sdc http://paste.ubuntu.com/10784781/ [18:43] bmorriso: the order the disks are picked up by the kernel are not persistent [18:43] bmorriso: so /dev/sdb would have been /dev/sdc in the installer environment [18:43] OK. Fair enough. It did what I wanted, which was to _not_ install on the local disks [18:44] bmorriso: you could use by-id path if you want a specific disk and you know its id [18:44] bmorriso: ls -l /dev/disks/by-id/ [18:44] bmorriso: that will make sure its that exact disk [18:45] this should suffice. I don't hard code paths, the only intention was to ensure it didn't install on the local disks. This is perfect! Thanks for the support. Love this channel/group/project [19:00] blake_r: what is the equivalent of "install disk-detect/multipath/enable=true" in curtin? Or is there? [19:18] I have this in curtin_userdata_ubuntu_amd64_generic_trusty_bfs2 http://paste.ubuntu.com/10785028/ but it doesn't actually install the package. [19:56] Would I not see early or late commands in install.log? Is there a way to see this output? [19:58] bmorris: try adding .local to the end of that file to see if it gets installed [19:59] might not be getting used [20:00] hostname will contain a .local if it was enlisted and not manually added [20:00] oh I know what's wrong [20:00] you need that one late_commands [20:01] and you need to use curtin in-target -- before it [20:01] without that you are installjng in the booted Ubuntu installed [20:40] Bug #1442360 was opened: 1.8b1: Cannot add machine Missing Power Type [21:01] blake_r: what you sent earlier, is now getting picked up but it errors http://paste.ubuntu.com/10785782/ [21:04] bmorriso: ah its block-meta not block_meta, my bad [21:05] yeah, I looked closer at the output...no worries. Appreciate the guidance. Wish there were more documentation on curtin. Feels like a black box [21:05] Is here the best place to ask curtin questions or is there a curtin specific channel? [21:06] bmorriso: this is the best place, no specific curtin channel [21:06] I just hacked together a pretty awesome preseed, and now I have to figure out how to get it all into Curtin. [21:30] So my install log shows "Installation finished. No error reported" but the server is just stuck in 'deploying' state http://paste.ubuntu.com/10785977/ [21:30] It never changes from 'deploying' to 'deployed' [21:36] bmorriso: the node has to boot again into the new installation and contactaas to [21:36] maas to be marked deployed [21:36] check the console of the node to see where it got stuck [21:37] It's at the login prompt [21:37] it never rebooted [21:37] It put the keys on the host [21:39] Here is a screenshot http://i.imgur.com/qXCk6Fb.png [21:42] I can provide cloud-init.log or cloud-init-output.log if it would be useful [22:22] Bug #1442379 was opened: 1.8beta1: 500 error when shutting down node