Adrian Chadd wrote:
2009/8/4 Hery Setiawan <yellowhat89@xxxxxxxxx>:
maybe in his mind (and my mind too actually), with big mem_cache the
file transfer will be transferred faster. But using that big for me
it's too much, since I only have 4GB of RAM and thousand workstation
connect with my squid.
The squid memory cache doesn't work the way people seem to think it
does. Once objects leave the memory cache pool they're out for good.
The rule of thumb is quite simple - keep cache_mem large enough to
handle in-transit objects and a few hot objects; leave the rest to be
available for general operating system disk buffer caching. Only
deviate from this if you absolutely, positively require it.
This will occur when your workload has a lot of small objects which
you frequently hit. Hack up or download something to generate a
request size vs {hit rate, byte hit rate, service time, cumulative
traffic} to see exactly how many tiny/small objects you're getting a
hit off of.
I recommend scalar.awk for this (http://scalar.risk.az/). Sample report
segment (which hopefully shows up in a fixed-width font...):
~~~ Objects Size Report
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
---------|------ R E Q U E S T S -------|----------- T R A F F I C
-----------
SIZE | total | misses | hits |hit%| total | misses |
hits |hit%
---------|--------|--------|--------|----|----------|----------|----------|----
0-0.1KB 19.426K 18.505K 921 5% 249.893 KB 249.893 KB 0
B 0%
0.1-1.0KB 248.507K 172.422K 76.085K 31% 123.209 MB 93.616 MB 29.593
MB 24%
1-5 KB 176.308K 156.113K 20.195K 11% 413.161 MB 364.980 MB 48.181
MB 12%
5-10 KB 51.447K 46.284K 5.163K 10% 352.742 MB 316.625 MB 36.117
MB 10%
10-50 KB 58.159K 49.855K 8.304K 14% 1.203 GB 1.032 GB 174.937
MB 14%
50-100KB 8.801K 7.656K 1.145K 13% 594.920 MB 516.522 MB 78.398
MB 13%
100-500KB 6.747K 5.843K 904 13% 1.216 GB 1.066 GB 153.852
MB 12%
0.5-1.0MB 710 654 56 8% 478.407 MB 440.962 MB 37.446
MB 8%
1-5 MB 558 498 60 11% 1.234 GB 1.105 GB 132.114
MB 10%
5-10 MB 137 126 11 8% 0.938 GB 883.506 MB 77.182
MB 8%
10-50 MB 96 73 23 24% 2.109 GB 1.417 GB 708.504
MB 33%
50-100MB 13 9 4 31% 813.873 MB 561.792 MB 252.082
MB 31%
>100MB 14 12 2 14% 4.650 GB 3.766 GB 905.610
MB 19%
===============================================================================
If you have a very small set of constantly "hot" traffic that will fit
in memory, up cache_mem. But be aware of the performance repercussions
if the "hot" traffic leaves cache_mem and stays on disk.. :)
If you have a set of "hot" traffic that moves over time, upping
cache_mem may not help.
2c,
Adrian
Chris
