> > I'm currently running Squid 2.6 as a reverse proxy server to ease the > load on my application web servers. > > The problem that I'm running in to is that Squid keeps using up all > available hard-disk and then stops servicing requests. > > My config looks like this: > > --------------- > http_port 80 accel defaultsite=app2.tveyes.com > > cache_peer app1.tveyes.com parent 8080 0 no-query originserver > round-robin weight=1 > cache_peer app2.tveyes.com parent 8080 0 no-query originserver > round-robin weight=1 > > visible_hostname app1.tveyes.com > > cache_dir ufs /var/spool/squid 2048 16 256 > > cache_mem 512 MB > > [snip] > > cache_log none > cache_store_log none > > emulate_httpd_log on > ---------------- > > I'm not sure why squid should be taking any more than 2GB of disk space, > but every time I look at the cache when it has crashed, the swap.state > file is 25GB. > > I apologize if this topic has been covered before, I've looked around > and not really found a satisfactory answer to why Squid is behaving this > way. You need to rotate your squid logs frequently, possibly daily. swap.state is a append-only journal and needs squid -k rotate to clean it up. Amos
