[00:04] <SpamapS> benburkert: I'd be interested in what that diff is :)
[00:06] <benburkert> SpamapS: https://gist.github.com/367891f799fc5769f5c6
[00:07] <SpamapS> benburkert: ok so datadir was moved
[00:07] <benburkert> right
[00:07] <benburkert> is it trying to use the old data dir maybe?
[00:07] <SpamapS> benburkert: no
[00:07] <SpamapS> benburkert: unless my.cnf tells it to
[00:08] <SpamapS> benburkert: if you stop/start the job, then do 'status mysql' does it say start/running ?
[00:15] <benburkert> SpamapS: it does start properly if i do that
[00:15] <benburkert> the problem is that i have chef scripts that think mysql is running because "start mysql" exits normally
[00:15] <benburkert> but it aint
[00:19] <SpamapS> benburkert: AHHH
[00:19] <SpamapS> benburkert: so later on, mysqld dies, and upstart does not respawn it?
[00:20] <SpamapS> benburkert: ultimately, there needs to be a chef module that asserts the state it wants (started, or stopped) and then uses status/start/stop to make that happen. I think the ones I've seen make invalid assumptions about the exit codes of start/stop/status
[09:05] <TauPan> Hi... I have a job that should start an agetty on ttyS*, as soon as it appears, but I have to start it manually. I wonder what's wrong with: 'start on device-added SUBSYSTEM=tty DEVPATH=ttyS*'
[09:08] <TauPan> it looks like the device is added pretty early, so I guess it might already be present as soon as upstart sees that 'start' line
[09:11] <TauPan> hm, maybe 'start on device-added SUBSYSTEM=tty DEVPATH=ttyS* or runlevel [2]' is better
[23:20] <SpamapS> TauPan: there is no "device-added" event. You probably want tty-device-added, per the explanation in the manpage of upstart-udev-bridge