[07:00] <[diablo]> Good morning #upstart ... guys how can I tell an upstart conf file not to fork please? I want the script thats getting called to fully run before upstart continues
[07:00] <[diablo]> or at least finish before the login prompt
[07:04] <[diablo]> would running the script as pre-start be better?
[15:31] <afournier> hi i updated an home made system from 1.6.1 to 1.7 so upstart runs as PID 1 and assert after processing some events:  "com.ubuntu.Upstart.c:2651: Assertion failed in control_emit_event_emitted: env != NULL"
[15:31] <afournier> any idea ?
[15:33] <afournier> i traced this problem up to this bug report, but here we are using it as init (PID 1)... https://bugs.launchpad.net/upstart/+bug/1161329
[18:20] <jeffreylevesque> any fine wizards here?
[18:20] <jeffreylevesque> i created an upstart script, which works.  Except, the datetime is always the same - https://github.com/jeff1evesque/machine-learning/blob/master/puppet/environment/development/modules/compiler/templates/webcompilers.erb#L38
[18:20] <jeffreylevesque> how can i adjust the `echo` statement, so the datetime correctly reflects the current time for the `echo`?
[18:30] <jeffreylevesque> any bots in here with enough ai to answer my q
[19:26] <chras> jeffreylevesque: whats happening now ? that when you do a start you get the same time ?
[19:27] <jeffreylevesque> had to move it into the `exec` - https://github.com/jeff1evesque/machine-learning/blob/3277f30c1bf47f84c449370c7d1dffdc3bccb084/puppet/environment/development/modules/compiler/scripts/sass#L14
[19:27] <jeffreylevesque> it's all fixed
[19:28] <chras> jeffreylevesque: yeah, thoes stanzas are their own independant sh's 
[19:28] <chras> cant pass variables between them
[21:18] <[diablo]> evening guys
[21:18] <[diablo]> can anyone please tell me how the hell I tell an upstart conf file to run a script that sleeps for 60 seconds, but does not fork... I want the entire boot process to pause for the full 60 seconds before continuing 
[21:19] <[diablo]> sadly I'm needing to do it in RHEL6 which is running upstart 0.6.5
[21:44] <chras> so you want rest of bootup to depend on that one sleeping event, just to pause bootup ?
[21:44] <[diablo]> hi chras 
[21:44] <[diablo]> well actually I'll explain
[21:44] <[diablo]> I have done for RHEL7 an expanding storage system with vmware guest ... 
[21:45] <[diablo]> on boot, it must rescan busses, expand the lvm, and grow the fs
[21:45] <[diablo]> problem is with RHEL6, the extending of ext4 takes some time... I don't want the boot to continue and the user to login before it's finished
[21:46] <[diablo]> I thought that the process was getting killed by upstart for hanging around too long... because I saw sometimes it only grew the fs say by half...
[21:47] <[diablo]> so I need really two things... to be sure that the boot waits until the grow fs has completed, and that the process does not time out
[21:47] <chras> while its off topic, couldnt you do thoes things concurrently while people are logged in ?
[21:48] <[diablo]> hmmm
[21:48] <chras> thoes are all safe to do while online
[21:48] <[diablo]> yes
[21:48] <[diablo]> but
[21:49] <[diablo]> if a user is in a rush, hypothetically he could login, not see the fs has fully expanded and wonder why it's not the size he'd asked for
[21:49] <[diablo]> I'd much prefer the delay being on the boot until login, and the fs fully expanded by the time he's logging in
[21:50] <chras> well you could always make your getty/gdm/whatever depend on your resize event finishing succesfully
[21:50] <[diablo]> yes 
[21:50] <[diablo]> thats true...
[21:50] <[diablo]> however I prefer not to edit those files... ideally I just want to drop in a /etc/init/resize.conf
[21:50] <[diablo]> and not mod anything else
[21:51] <chras> there shouldnt be any limitation to how long it takes for your upstart event to finish
[21:51] <[diablo]> chras, yeah actually I think the timeout was when I was trying out doing pre-script or something akin
[21:51] <chras> you could temporarily overwrite the start on for your getty.conf
[21:51] <chras> with getty.override
[21:52] <[diablo]> oh ok
[21:52] <[diablo]> one sec. let me find something
[21:52] <chras> i havnt done any long running pre-start's
[21:52] <[diablo]> guess it's minigetty on rhel
[21:53] <chras> but i quite often do running-forever stuff in script stanaza's 
[21:53] <[diablo]> oh I've got a starttys
[21:54] <chras> yeah i dont know what the redhat upstart events are, or if you are using custom ones
[21:55] <[diablo]> chras, can't find much info really on upstart on rhel6, just the norm
[21:55] <chras> i use all custom events for our events
[21:55] <chras> never used any standard or base ones
[21:56] <[diablo]> I'll look into this override 
[21:56] <chras> that might be simpliest way to do waht you want
[21:56] <chras> i think you need 1.3+ for it
[21:57] <[diablo]> doh
[21:57] <[diablo]> no, it supports it 0.6.5
[21:57] <[diablo]> just found it in the man
[21:58] <[diablo]> so I'd put something like start on resize
[21:59] <chras> start on stopping resize 
[21:59] <chras> best bet
[21:59] <chras> since you want it to trigger after your resize event is finished
[21:59] <[diablo]> ah gotcha
[21:59] <[diablo]> giving it a whirl now
[22:00] <chras> and then you could even flag that based on succes/fail
[22:00] <chras> but i assume in the end, you want the user to be able to log in regardless
[22:02] <[diablo]> hmmm sadly not working, most likely my mistake...
[22:02] <[diablo]> but you've given me food for thought
[22:02] <[diablo]> many thanks chras 
[22:03] <chras> what didnt work ?
[22:03] <chras> didnt get your getty?
[22:03] <[diablo]> did't wait for login prompt
[22:04] <chras> so your resize ran , and then your login prompt fired
[22:04] <chras> without resize finishing
[22:04] <[diablo]> ah I see I did something wrong
[22:04] <[diablo]> one sec
[22:06] <[diablo]> hahaha chras I think you've nailed it... I'm running a sleep 60 cos I'm testing it out at home on a vm
[22:06] <[diablo]> splash has gone, no login yet...
[22:07] <[diablo]> chras, dude u da man
[22:07] <[diablo]> works!
[22:07] <chras> nice one
[22:07] <[diablo]> thats exactly what I wanted !
[22:07] <chras> you might have to check a fail case with that, just in case a resize never happens
[22:07] <[diablo]> chras, aye
[22:08] <[diablo]> dude many thanks really
[22:08] <chras> are you doing your resize in a script, or a pre-start script?
[22:08] <[diablo]> it's been pissing me off all afternoon...
[22:08] <[diablo]> now it's gone midnight here ... gonna have to head to bed
[22:08] <chras> i dont recall if a stopping event will fire if a pre-start fails
[22:08] <[diablo]> cheers dude and all the best! god bless IRC :)
[22:08] <chras> have a good one
[23:05] <mjnowak> Anyone have a minute to possibly help me out with a nginx not respawning problem?