On 16/09/22 15:26, Grant Taylor wrote:
On 9/15/22 8:29 AM, Hildegard Meier wrote:
Hello,
Hi,
we have two Squid servers (Linux hosts) and each shall have the very
same config file /etc/squid/squid.conf which is versioned and deployed
from a central deployment server. So each host shall have deployed
the same files.
Each of the two shall have the other configured as sibling cache peer.
So node1 shall have
cache_peer node2.examlpe.com sibling 3128 3130
and node2 shalle have
cache_peer node1.examlpe.com sibling 3128 3130
I have no idea if it will work or not, but I might be tempted to try --
what I first saw as -- Solaris's "mailhost" or "losthost" name overloading.
Meaning that you would have a single identical entry in the config file
on both systems:
cache_peer peer-cache-node.example.com sibling 3128 3130
And then each host would use /etc/hosts to resolve the
peer-cache-node.example.com name to it's neighbor's IP. E.g.
node1:/etc/hosts
...
192.0.2.1 node1.example.com
192.0.2.2 node2.example.com peer-cache-node.example.com
...
node2:/etc/hosts
...
192.0.2.1 node1.example.com peer-cache-node.example.com
192.0.2.2 node2.example.com
...
I sort of suspect that the nodes already have a per-node /etc/hosts file.
This is untested and wild speculation on my part. Can / will someone
with more experience comment on the viability of such "mailhost" or
"loghost" use?
This config does work.
Although, I personally would first look at the system being used to
distribute the config files. All the ones I am aware of some sort of
macro substitution in files as they are pushed out for exactly this purpose.
Alternatively the cache_peer config settings can be pushed as a separate
file in /etc/squid/conf.d/cache_peer.conf and the main config file use
'include' directive to load it.
Cheers
Amos
_______________________________________________
squid-users mailing list
squid-users@xxxxxxxxxxxxxxxxxxxxx
http://lists.squid-cache.org/listinfo/squid-users