question (used to be all set to 1 second worth
of traffic). Pools are restored to this level
on reconfiguratoin.
+ - Changed storeClientCopy to give a swap-in failure if
+ the number of open disk FD's is above the 'max_open_disk_fds'
+ limit. Otherwise, a very loaded cache will end up with
+ all disk files open for reading, and none for writing.
Changes to Squid-2.1 (November 16, 1998):