> I, being a mainframe dinosaur, just wrote a script that checks for the
> processes' existence.  If it ain't there, it starts it. I cron it
every 5
> to 15 mins based on how important the task is.  (Actually, it does a
> netstat and checks for the port.) Samples available for the asking.

Yeah, implementing the "probe" method in the init script would be the
way to make that work as a systematic approach, but the disadvantages to
this method are:

1) It relies on cron to be functioning. If a system has hit resource
exhaustion and has started killing processes to try to stay alive, cron
may be a early victim and what's going to restart cron if it dies?

2) There's the delay between the death of the process and the cron check
for it. The spawner idea will react as soon as the kernel processes the
SIGCHILD for the spawned process.

----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390

Reply via email to