[15:04] <adrien_oww> hi, is there a tool that can give me a pretty graph of when a service starts, stops, ... ?
[15:04] <adrien_oww> I'm doing boot-speed optimization on a fairly big system and bootchart shows I have two big bottlenecks which are almost certainly due to service delays (like "sleep 30"...)
[15:05] <adrien_oww> but I can't tell which ones
[15:07] <adrien_oww> otherwise I'll write my own :) 
[15:19] <jodh> adrien_oww: not currently - the best option atm is to create a job that calls 'upstart-monitor --no-gui' as that will log all events (including job starts/stops) along with high-res timestamps. Then just munge the output using gnuplot/blah.
[15:21] <jodh> adrien_oww: you could just write your own script though since upstart-monitor is simply listening for the EventEmitted D-Bus signals.
[15:38] <adrien_oww> ok, thanks :) 
[15:38] <joelteon> "start mysql" is producing no output for me and exiting 0. it also doesn't start mysql
[17:33] <Curly060> Hi! I am using upstart to run nginx as webserver. This works fine until I want to use nginx's possibility to update the binary on the fly. This is done by sending a signal to nginx. In that case nginx spawns a 2nd process and phases out the old one. It also updates the PID file. However, upstart is now confused after this. status shows, that nginx stll runs, but it shows no pid. stop says, the job was stopped, but nginx still ru
[17:33] <Curly060> ns. start claims everyting is ok, except it shows no pid. reload says "unknown instance".
[17:35] <Curly060> using the old start-stop-daemon functionality the whole procedure is not a problem. But my guess is that upstart works entirely different, monitoring the provess directly and the above procedure "steals" the process which upstart monitors. maybe upstart even tries to relaunch the process, but fails, because nginx is running still.
[17:35] <Curly060> So is there any way to use upstart and the described change-binary-on-the-fly method with upstart?
[21:05] <tty-1> hi
[21:05] <tty-1> anyone online? :)