[03:07] <_Andrew> Hi [03:08] <_Andrew> I have an upstart script I have made but it's not working on startup http://paste.ubuntu.com/351057/ [03:08] <_Andrew> Am I doing this right? I'm trying to execute a program under a specific user [04:24] _Andrew: 'exec' has a '-u' option on your system ? [04:25] <_Andrew> I just copied it from so other upstart script [04:25] <_Andrew> I have no idea [04:25] from which one? [04:26] you'll need to use something like su or sudo IMO [04:29] unless you do have an 'exec' with that option of course [04:30] <_Andrew> http://wiki.dropbox.com/TipsAndTricks/TextBasedLinuxInstall?highlight=%28upstart%29 [04:30] <_Andrew> I got it from here [04:31] well, they *do* use sudo there, right? [04:33] also, the 4 first lines can be written as one: start on runlevel [2345] [04:34] Also, [-f foo] is invalid (unless you have a command called [-f), you’ll want [ -f foo ] [04:35] Will startup.sh stay in the foreground until shutdown.sh is called? [04:36] (The job you pasted expects it to.) [04:36] <_Andrew> I think so yes [04:38] <_Andrew> Should it not stay in the foreground? [04:39] It should. [04:40] To clarify: it should stay in the foreground. :-) [04:40] ion: well, that or you need extra stuff in that file ;) [04:40] A future version of Upstart will properly track daemonizing things, but for now, it’s often best just to have it stay in the foreground. [04:41] janc: Yeah, and you also need to know exactly what you’re doing in order not to break Upstart’s current fork-tracking code. :-) [04:41] _Andrew: did you read init(5) ? [04:43] http://manpages.ubuntu.com/manpages/karmic/en/man5/init.5.html for a web version [04:44] it's the best docs upstart has currently, I think ;) [04:44] <_Andrew> ok thanks gys [04:45] _Andrew: I also see that dropbox example is for versions of Ubuntu < 9.10 [04:46] the script needs to go into /etc/init/ now, not /etc/event.d/ [04:46] so, depending on what version of upstart you use... [04:47] <_Andrew> Our server is LTS so I think it's in the right place, but thanks for letting me know [04:47] _Andrew: ah, then the manpage I pointing to might not work [04:48] I'm not sure if there is much documentation for that version of upstart [04:48] _Andrew: you can still use sysvinit init scripts though, if you are more familiar with those [04:49] <_Andrew> I think this should be working now. I'll let you know if I have any other probs that I can't figure out === ev1 is now known as ev [13:44] keybuk: O hai [13:44] Welcome back [13:44] keybuk: http://people.canonical.com/~scott/daily-bootcharts/ :-D [13:44] Happy new tiger [13:44] Keybuk: can you remind me the current plan wrt upstart and network up/down events? [13:44] ion: what about them? [13:45] ion: ah, amusing gap for new year calcuations :p [13:45] keybuk: Nothing really. It was just interesting to learn there are 29 days in the zeroth month of the year. [13:45] Beginning with the 0th day, of course. [13:45] Md: I don't have any current plans fori t [13:46] Keybuk: I do... what do you think about a long-lived daemon which will listen to interface events on the netlink socket and feed back "cleaned up" events to upstart after applying a "carrier delay" and/or debouncing? [13:47] Md: that's probably the right kind of thing [13:47] though I don't know what you mean by "carrier delay" [13:48] oh, it's an IOS feature which delays up/down events for a few ms, to prevent flapping the IGP if the physical layer bounces === sadmac2_ is now known as sadmac2 === robbiew_ is now known as robbiew === robbiew is now known as robbiew-AF === robbiew-AF is now known as robbiew-afk === robbiew-afk is now known as robbiew === robbiew is now known as robbiew_ [23:55] Md: bonding uses similar delays to deal with network equipment bring a link up before it is ready to recieve and pass traffic, which is another applicable analogy IRT upstart.