[14:58] <pmjdebruijn> hi all
[14:58] <pmjdebruijn> somehow my raid isn't being properly assembled, I get a /dev/md127
[14:59] <pmjdebruijn> it is possible that the config isn't present yet, since we do a netboot, which needs to sync the configs first
[14:59] <pmjdebruijn> the problem is as I've had before dependancies between upstart jobs, and sysv jobs
[14:59] <pmjdebruijn> I'm wondering whether converting the mdadm sysv script would be a good idea
[15:01] <pmjdebruijn> I basically need the mdadm script to depend on a custom upstart job
[15:18] <dcorbin_work> Can I call exit from an upstart script safely?  Do the values indicate success/failure?
[15:20] <jhunt> pmjdebruijn: if you're running ubuntu, you might want to raise a bug with full details.
[15:21] <jhunt> dcorbin_work: yes. See http://upstart.ubuntu.com/cookbook/#standard-environment-variables (see EXIT_STATUS)
[15:21] <dcorbin_work> jhunt: thanks
[16:31] <dcorbin_work> jhunt: That documents what upstart sets.   Not what my upstart script should do to clearly indicate failure. (at least, that's how I see it)
[16:42] <savid> How can I emulate start-stop-daemon's "-c <user>:<group>" option in upstart?
[16:43] <jhunt> dcorbin_work: 0 success, anything else is failure unless you override it: "man 5 init" and search for "normal exit".
[16:44] <jhunt> dcorbin_work: normal exit is for respawning of services though.
[16:45] <dcorbin_work> I think I've found another way to do what I want, but what I was trying to do was "check for a pre-condition", and fail if it wasn't there.
[17:12] <dcorbin_work> job A has "start on starting B".  I want to be sure A is completely up before B starts.  If I have a post-start script that checks to make sure A is completely up, is that adequate?  Will a start request for B wait until all processors of "on starting B" have finished?
[18:22] <dcorbin_work> How can I see all the events that init is triggering?
[18:36] <jhunt> dcorbin_work: your job A/B example doesn't make sense - which job starts first? "start on started X" (not "starting") might do what you want?
[18:38] <dcorbin_work> jhunt: for B to run, A must be up.  But, just because A is started, doesn't mean B should start
[18:43] <jhunt> still not clear exactly what you're trying to do. I have to go now, but if please feel free to post further details to the upstart-devel mailing list at the link in the topic.
[18:43] <jhunt> In answer to one of your other questions: http://upstart.ubuntu.com/cookbook/#obtaining-a-list-of-events
[18:57] <savid> Does anyone know how can I emulate start-stop-daemon's "-c <user>:<group>" option in upstart?
[18:58] <savid> basically, change the user and group of the running process?
[22:06] <hallyn_> interesting - so pre-start stanzas only run when there's a job to actually kill?  if you have a job with only a pre-start and pre-stop, then pre-stop does nothing...
[22:06] <hallyn_> SpamapS: ^ perhaps that should be 
[22:06] <hallyn_> mentioned in the cookbook?