All, I'm interested in using Squid in an accelerator configuration in front of http://mirror.anl.gov/. We serve many ISO images, and I'm interested in decreasing the load on our storage subsystem. I'm interested in configuration recommendations. Here are two links (generated twice an hour) that shows a list of all files currently being downloaded from our site for each of our two front-end machines. These links give a good idea of the kind and quantity of files I'd like to accelerate: http://alpaca-i.it.anl.gov/files http://alpaca-ii.it.anl.gov/files The hardware in use is described here: http://mirror.anl.gov/#farm I have a tentative squid configuration that includes two local independent 72G disks for cache on each machine, and the following relevant details. I'm pretty sure I'm way off on the object size stuff: cache_mem 1000 MB maximum_object_size 5120000 KB maximum_object_size_in_memory 1024000 KB cache_dir aufs /var/spool/squid/cache1 51200 16 256 cache_dir aufs /var/spool/squid/cache2 51200 16 256 httpd_accel_host localhost httpd_accel_single_host on httpd_accel_uses_host_header on Thanks! -Brian -- Brian Elliott Finley Linux Strategist, CIS Desk: 630.252.4742 Cell: 630.631.6621
