Depends on what you mean by crash. If you mean the daemon core dumps,
then you should send in a bug report so the authors can figure out why.
Beyond that, write a little shell or perl script to check the process
table and start squid if it isn't running. Have it run out of cron every
five minutes or so.
If you mean, in case of the box crashing, it should restart as long as
there is a startup script for it as soon as the box comes back up. If the
box goes down and you don't want to drive out to fix it, you can always
have someone power cycle it. WARNING: Power cycling a unix box with
mounted filesystems is VERY BAD. While 98% of the time, you'll get away
with it and it will fix itself on reboot, every once in awhile the fs will
be seriously trashed and manual intervention will be necessary to get the
box to boot again. If you have a user who is not completely braindead,
you could create a reboot account that has shutdown as it's shell and have
the user "login" as reboot, wait for the box to shut down, and then power
cycle.
Just a thought.
On Mon, 17 Apr 2000, Jimmy Stewpot wrote:
> Hello all,
> If we where to have a crash, is there anyway that we can get squid to automatically restart, or anything like that apon a major errors. This is because we have several squid servers around the country and would prefer not to have to drive out to restart as we have had to stop telnet/ssh access.
> thankyou
> Jimmy Stewpot
>
>
Received on Sun Apr 16 2000 - 23:52:09 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:52:57 MST