I ran into the same problem when first setting my squids up. My solution was
to set up proxy.greenapple.com as the box all browsers are pointed to, but
only us a 200MB cache dir and very little cache_mem. This has keep the
machine from every going down (knock on wood) and then set up all other
squid as proxy-only parents to this box with the big HD and mem. Thus when
request comes into proxy.greenapple.com, it asks all parents and the parents
do all the work. Hence if a parent, or even all the parents, crash, the
browsers(customer) still gets the pages.
Mark
----- Original Message -----
From: Gene Black <gblack@3wave.com>
To: <squid-users@ircache.net>
Sent: Friday, September 17, 1999 11:08 AM
Subject: Hot Standby
> Since squid tends to become a single point of failure, what methods are
> commonly utilized to automatically deal with it if it crashes? I'm
> thinking of something along the lines of Cisco's Hot Standby Router
> Protocol or the such, though, to the best of my knowledge, there is no
> Linux support for such an item.
> I have sevral items that use the squid server as their default gateway
> (making it easy to do transparent proxying), and if/when the squid box
> crashes, I'd like for something else to step in to at least route the
> traffic, if not provide a backup proxy. I was wondering if there were
> any utilities that would monitor an IP address and then bring that IP
> address up on it's host machine if it couldn't reach the IP it was
> monitoring for X amount of time (and maybe even detect when the device
> comes back up and then drop the IP address on it's host).
>
>
> Thanks in advance,
>
>
> Gene
>
>
Received on Fri Sep 17 1999 - 14:23:47 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:48:28 MST