On Wednesday 23 November 2005 01:40, John Cougar wrote:
> Info,
>
> Are you looking to slow squid down to a grinding halt?? ACLs are in memory
> for a very good reason, and forcing Squid to go to Disk I/O for every access
> would give you ... less than desirable results, rest assured.
OSes started to cache file data in RAM many years ago:
# cat t.c
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
char buf[64*1024];
int main() {
int i = 1000000;
while (--i) {
int fd = open("t.c", O_RDONLY);
if (fd<0) return 1;
read(fd, buf, sizeof(buf));
close(fd);
}
return 0;
}
# gcc -Os t.c
# time ./a.out
real 0m7.057s
user 0m0.640s
sys 0m6.260s
In other words: my Celeron 1200 MHz just did more than 100000 open/read/close's
per 1 sec.
--
vda
