[01:41] kreignf: are the modules not loading, or are they loading too late? [01:42] remember that many upstart jobs run in parallel [01:59] Am I right in thinking that neither --verbose nor --debug show me the FOO=bar values associated with each event? [02:00] And if so, is there a way to see those values? [03:11] SpamapS, I will say this... dear god it's awesome having a system boot 'completely' off of USB before I can walk the 5 feet to my keyboard [03:11] ... and log in remotely [03:26] kreign: you using live-boot? [03:27] twb, I'm not sure what live-boot is. [03:27] twb, though it sounds interesting; what is it? :) [03:27] It's a fork of casper [03:27] With the idea of being useful for everybody, rather than just canonical [03:36] twb, hmm no, I'm not all that familiar with it, to be honest. I was just trying to get upstart to work with zfs. [03:36] on ubuntu [03:36] zfs-fuse? [03:36] no [03:36] * twb blinks [03:36] There's an Ubuntu kFreeBSD now? [03:36] no [03:37] it's ubuntu with the CDDL ported modules. [03:37] just ubuntu [03:37] takes about 10 minutes. [03:37] ... just have to figure out the little 'ifs' and 'buts' :P [03:37] I guess that's OK if you don't redistribute the result... [03:37] no [03:37] but the knowledge of how to do so, or the deb packages, should be fair game. [03:38] twb, if you're curious I'll stick in here tomorrow and give you the info [03:38] time to put the kids to bed though [03:38] Not really; my money's on btrfs [03:38] eh [03:38] have you used zfs? :P [03:38] Yes, on osol. [03:38] ah [03:38] imo btrfs is years behind [03:38] Granted. [03:39] I don't have time to wait around. [03:39] for what I need, zfs is the trick [03:39] stable or no [03:39] (and it appears to be acceptably so, fwiw... I've not been able to break it yet, and the freebsd implementation is somewhat easy to cause issue with) [03:39] at any rate [03:39] * kreign poofs [03:40] But CDDL means that Debian can't ever distribute systems that support ZFS out of the box, so I'm not inclined to jump through hoops for it. [03:40] It's the same reason I avoid MP3 [05:05] kreign: there was somebody with the same problem last week or so [05:05] twb: MP3 has an expiration date though [05:05] SpamapS: granted. [05:06] kreign: and like I suggested before, your problem is that the module gets loaded too late [05:06] if wikipedia is to believe, decoding may be patent free in the US just about the time the world ends, December 2012 [05:07] kreign: you need to make sure all zfs modules are loaded before trying to mount the ZFS filesystem... [05:07] ahh yes [05:07] startup and started udev will race with mountall [05:08] JanC: how do you do that? Do you have to patch the mountall script to say start on [...] && modules? [05:08] change it to 'start on mounting TYPE=zfs' [05:08] mountall blocks on mounting and mounted events [05:08] they are considered hook points [05:09] its tricky, but if you look at the way nfs-ustils was recently patched in natty, you can actually use them in an and-like fashion with other events. [05:10] ustils.. my typing has really suffered since losting my old MS elite keyboard.. need to go find another one. :-/ [05:10] I think a "task" that loads the zfs modules, and has "start on starting mountall" or something like that might suffice, but I would have to look into the logs what he finally used [05:11] well, anything that makes mountall wait on that task to finish ;) [05:13] starting mountall might work too [05:14] Actually yeah, since modules will always be on the rootfs, I think thats a nice simple solution. [05:23] SpamapS: no model M? [11:08] JanC: ping [15:49] JanC, for whatever reason, including the modules in /etc/modules does not result in them being loaded. not sure why. [15:50] JanC, also, I was that person with the problem last week. I never got it resolved then, though I figured out a workaround. :)