[squid-users] Squid using up all Swap Space

Squid 2.4.STABLE1 seems to work well for several days and then comes to a
crawl. I have 768MB RAM in the machine and after about 4-8 days (of very
little swapping), I then see continuous swapping and I eventually run out of
swap space (I have 256MB set up as swap partition). I've been getting
thousands of these errors below in my cache.log when experiencing this
problem. Can someone help me determine what exactly is happening and how I
fix it? Thanks.

2001/04/22 16:56:46| httpAccept: FD 12: accept failure: (104) Connection
reset by peer
2001/04/22 16:56:50| comm_accept: FD 12: (104) Connection reset by peer
2001/04/22 16:56:50| httpAccept: FD 12: accept failure: (104) Connection
reset by peer
2001/04/22 16:56:51| comm_accept: FD 12: (104) Connection reset by peer
2001/04/22 16:56:51| httpAccept: FD 12: accept failure: (104) Connection
reset by peer
2001/04/22 16:56:52| comm_accept: FD 12: (104) Connection reset by peer
2001/04/22 16:56:52| httpAccept: FD 12: accept failure: (104) Connection
reset by peer
2001/04/22 16:56:52| comm_accept: FD 12: (104) Connection reset by peer
2001/04/22 16:56:52| httpAccept: FD 12: accept failure: (104) Connection
reset by peer
2001/04/22 16:56:54| comm_accept: FD 12: (104) Connection reset by peer
2001/04/22 16:56:54| httpAccept: FD 12: accept failure: (104) Connection
reset by peer
2001/04/22 16:56:55| comm_accept: FD 12: (104) Connection reset by peer

Additional Squid configuration:

cache_dir ufs /cache2 35000 32 256
cache_dir ufs /cache3 35000 32 256
cache_mem 32 MB
maximum_object_size 10240 KB
maximum_object_size_in_memory 1024 KB

I am also running this transparently with a Cisco router and WCCP.

Shane T. Ferguson
Received on Sun Apr 22 2001 - 14:15:37 MDT