[07:48] <fozze> ok, i did a very bad thing
[07:49] <fozze> i installed the upstart package and then rebooted without modifying my configuration on my vps
[07:49] <fozze> and now i am screwed
[07:50] <mbiebl> boot with init=/bin/bash
[07:50] <mbiebl> And the run /etc/init.d/rcS && /etc/init.d/rc 2
[07:53] <fozze> unfortunately i can only access the computer remotely
[07:55] <fozze> however, i found a way to temporarily boot up the computer in another mode that my host provides and it allows me to see what procs are running, and so far the only proc that is running is init
[07:57] <mbiebl> can you access the computer via serial line?
[07:58] <fozze> there is a "repair" mode that allows a simple working filesystem to boot up and then my old filesystem is placed in a directory called /repair so I can go in there and manually fix stuff, i just don't know what to edit to get upstart to work
[07:58] <fozze> so right now, i'm in repair mode
[07:58] <mbiebl> Which system do you run?
[07:58] <fozze> ubuntu edgy
[07:59] <fozze> it was running sysvinit
[07:59] <fozze> then i did a sudo apt-get install upstart and then rebooted
[07:59] <mbiebl> Do you have system-services/upstart-compat-sysv/system-tasks installed?
[07:59] <mbiebl> If not, do that.
[07:59] <fozze> hmm i don't know..
[07:59] <fozze> probably not
[08:01] <fozze> let's check
[08:02] <fozze> is there a way i could use dpkg to manually install those packages under a /repair folder?
[08:03] <mbiebl> chroot /repair
[08:03] <fozze> cool
[08:04] <mbiebl> btw. s/system-tasks/startup-tasks/ 
[08:04] <mbiebl> (not that it would matter, because the package doesn't contain any relevant files yet)
[08:05] <fozze> i see
[08:05] <fozze> i probably need to use the dpkg under /repair don't i...
[08:06] <mbiebl> after you have chrooted to /repair, apt-get install upstart-compat-sysv should work.
[08:07] <fozze> ah, it does :)
[08:08] <fozze> hmm
[08:08] <fozze> system-tasks, is that the correct name
[08:08] <fozze> ?
[08:08] <mbiebl> btw. s/system-tasks/startup-tasks/ 
[08:08] <mbiebl> quoting myself ;-)
[08:11] <fozze> so, after doing all this, it should use my old sysviconfig when i boot up normally?
[08:11] <fozze> sysvinit config rather
[08:12] <mbiebl> yeah, it use the sysv init scripts.
[08:12] <fozze> all right, rebooting...
[08:14] <fozze> hrmm
[08:14] <fozze> well i wasn't able to install system-tasks
[08:15] <fozze> and so it seems like process is right where it was before
[08:18] <mbiebl> s/system-tasks/startup-tasks/  means, substitute system-tasks with startup-tasks
[08:18] <mbiebl> I made a typo at the beginning
[08:18] <mbiebl> Have you checked that upstart-compat-sysv and system-services are installed properly?
[08:18] <fozze> i'm checking
[08:19] <fozze> yep, they are there
[08:20] <mbiebl> why did you have sysvinit at first place?
[08:20] <mbiebl> edgy ships with upstart by default
[08:21] <mbiebl> maybe you messed something up.
[08:21] <fozze> it came like that, that's what i thought
[08:21] <fozze> my host provider provided some whacked up version of ubuntu edgy
[08:21] <fozze> actually
[08:21] <mbiebl> Well, if they don't ship upstart, which is default, they must have a reason.
[08:21] <fozze> they are using virtuozzo
[08:22] <fozze> it may only work with sysvinit
[08:22] <mbiebl> Very likely, upstart does not work with virtuozzo yet.
[08:22] <mbiebl> Best is, you go back to sysvinit.
[08:23] <fozze> i just wanted my startup scripts to be simpler :)
[08:23] <mbiebl> Or you feel adventurous and help to debug the problem ;-)
[08:23] <fozze> lol
[08:23] <fozze> yes, with all my free time *sarcastic*
[08:23] <fozze> well thanks for your help?
[08:23] <fozze> help.
[08:24] <mbiebl> Please file a bug report at launchpad.net against upstart.
[08:24] <mbiebl> So this issue can be tracked.
[08:24] <mbiebl> Describe the hardware/software environment that you use.
[08:24] <mbiebl> etc.
[08:25] <mbiebl> https://bugs.launchpad.net/upstart
[08:38] <fozze> so you think it's a problem with upstart and not virtuozzo?
[08:38] <fozze> or both?
[08:39] <mbiebl> Hard to tell.
[08:39] <mbiebl> At least file the bug report so this discussion is not lost.
[08:51] <fozze> ok, well going back to sysvinit works...
[09:13] <fozze> ok, bug filed
[11:51] <Keybuk> heyhi
[11:53] <_ion> Hi
[11:54] <Keybuk> how goes it?
[11:55] <_ion> Well, i'm alive. :-)
[11:56] <_ion> How are you?
[11:56] <Keybuk> just back from the gym
[11:57] <Keybuk> been thinking about complex-event-config
[11:57] <Keybuk> and ironically, have decided it's not as simple as I though
[11:57] <Keybuk> +t
[11:59] <_ion> Are you thinking of redesigning it?
[12:01] <Keybuk> I mean in the way that it interacts
[12:01] <Keybuk> in particular with event emissions and instance jobs
[12:01] <Keybuk> take the simple example
[12:01] <Keybuk> on block-device-added and some-other-event
[12:01] <Keybuk> there's actually four or more things you might want that to mean
[12:02] <Keybuk> 1) whichever event causes the goal change (the second one to occur) is blocked until the job finishes, and is the one the job sees in its arguments and environment
[12:03] <Keybuk> 2) you want a specific event (e.g. the first) to be blocked and appear in the arguments; this might need some syntax ..."on using block-device-added and some-other-event"
[12:03] <_ion> True...
[12:03] <Keybuk> 3) you want a combination of all of the events to be blocked and appear in the arguments (but how would you get the arguments?)
[12:04] <Keybuk> I can't remember the 4th thing
[12:05] <_ion> on block-device-added as foo and some-other-event as bar  args: $foo_1, $foo_2, ..., env vars: $foo_DEVNAME, $bar_BAZ
[12:07] <_ion> on using block-device-added and some-other-event as bar  block-device-added: $1, $2, ..., $DEVNAME; some-other-event: $bar_1, $bar_2, ..., $bar_BAZ
[12:08] <Keybuk> would you block those events/
[12:08] <Keybuk> if so, what happens if block-device-added happens but some-other-event never happens ... the event would never be finished
[12:08] <_ion> I haven't thought of it yet.
[12:10] <_ion> Perhaps a syntax such as 'on block-device-added [block, using]  and some-other-event [as bar] '
[12:10] <Keybuk> gets tricky, doesn't it :-/
[12:11] <_ion> Indeed. :-)
[12:11] <Keybuk> not surprising really, since the model so far is assumed simple