Hi,
I am looking for return of experience on successful deployment of Squid on a Linux HA cluster.
My main question is : which strategy to adopt regarding the cached items ? Use a shared network drive ? Or in case of failure just start the slave proxy disregarding what is in its cache, and thus have a slightlty downgraded user experience.
Thank you.
Hi,
what kind of setup are you thinking about? Transparent proxy?
I wouldn't use a shared network drive as I/O Performance is critical for squid. If you use a loadbalancing/ha solution both proxies will have a recent data in the cache.
I will use proxy.pac-loadbalancing/ha when we deploy our second squid. Sounds nice & simple.
Regards, Hendrik Voigtländer
