Hi again!
I have 2 identically configured cache servers running as siblings to each
other serving ~ 4,500 clients together. The configurations are : Core 2
Duo 2.33 GHz, 8 GB RAM, 1 x 160 GB IDE (OS), 4 x 160 GB SATA 3.0 Gbps
(cache_dir).
My LRU Reference Age (in Store Directory page) is only between 3.00 and 3.41
days. If I read understood it correctly, it means most of my cache contents
are replaced in less than 3.41 days. Does anybody have any idea if this is
a low value or not?
I'm using only 65G on each HDD as cache_dir, which is roughly 40% of the
drive capacity. I read somewhere in the mailing list that we can go upto
80% of drive capacity safely. But when I do even 80G, my IOWait goes way
high. :( Does that mean I need faster HDDs?
My performance is reported in the benchmark:
http://wiki.squid-cache.org/KnowledgeBase/Benchmarks#head-99a7a6b698d2e97de2bdd4385cb423cd7c788bf8.
We are missing the "Byte Hit Ratio" there which stands at between 15 adn 20%
during peak hours.
The config file can be found here:
http://116.193.170.11/squid/squid_2008-11-21-1000_config.html. Some
highlights of the config are:
cache_mem 4294967296 bytes
maximum_object_size_in_memory 65536 bytes
memory_replacement_policy lru
cache_replacement_policy lru
cache_dir aufs /cachestore/cache1 65536 16 256
cache_dir aufs /cachestore/cache2 65536 16 256
cache_dir aufs /cachestore/cache3 65536 16 256
cache_dir aufs /cachestore/cache4 65536 16 256
minimum_object_size 0 bytes
maximum_object_size 1073741824 bytes
cache_swap_low 90
cache_swap_high 95
update_headers on
access_log /var/log/squid/access.log squid
cache_log /var/log/squid/cache.log
