On Wednesday 06 August 2003 17.58, SSCR Internet Admin wrote:
> yeah thats true. but IMOP, using null fs and holding those object
> in RAM "should" give a better squid performance or maybe a good hit
> rate since squid functions the same way as having a null fs or
> having a big cache_dir except no IO bound problems (since this is
> the hindrance).  I asked this coz i have seen on cachemgr that only
> have 10%-19% hit rate using null fs (with 250MB cache_mem)compared
> to 25%-45% using cache_dir (with 250MB cache_mem).

Makes sense. There is higher pressure on cache_mem than cache_dir, and 
even more so when you do not use a cache_dir (i.e. cache_dir null is 
a cludge to not have any cache_dir).

The cache_mem is not only used for cached objects, but also in-transit 
objects. Usually in-transit objects use only 16 KB per ongoing 
request but in some situations more may be used.

What you can try is using a ramdrive instead of a large cache_mem. 
This may provide a higher hit ratio/MB.

-- 
Donations welcome if you consider my Free Squid support helpful.
https://www.paypal.com/xclick/business=hno%40squid-cache.org

If you need commercial Squid support or cost effective Squid or
firewall appliances please refer to MARA Systems AB, Sweden
http://www.marasystems.com/, [EMAIL PROTECTED]

Reply via email to