Tim Lesher wrote:
Looks like Trac went down. TextDrive recently decided to stop using
the "daedalus" tool to keep processes running, and I'm not sure what
their replacement is...
I just use an hourly cronjob to start tracd. It fails gracefully if tracd is
already running.
Careful with that... the whole reason TxD banned Daedalus is because
of runaway Rails processes hanging, get restarted, hanging, getting
restarted, hanging, etc, which brought down an entire shared server on
several occasions.
I was reminded by this discussion of supervisor, which happens to also
have a timeout for this case, among other features:
http://plope.com/software/supervisor/ -- it's one of those things I've
meant to use for a long time, but haven't sat down and fully figured
out. Not that it seems particularly hard.
I'm not sure if it has all the specific features daedalus has. It
doesn't monitor processes. But if something else monitored and killed
hung processes, it would restart them. It could, for instance, take the
place of the reloader (which would still monitor, but would just die
with a non-0 exit code when there is a problem).
--
Ian Bicking / [EMAIL PROTECTED] / http://blog.ianbicking.org
