On 16.08 21:25, lokesh.khanna@xxxxxxxxxxxxxxxxx wrote: > I am using 2 Scsi hard disk of 36 GB for cache. I have assigned only 15 > GB ( less than 50 % of total disk ) for cache in each disk . > So total cache is 30 GB. > > cache_dir diskd /cache1/squid 15360 16 256 Q1=64 Q2=72 > cache_dir diskd /cache2/squid 15360 16 256 Q1=64 Q2=72 > > > I want to know if squid uses full 30 GB for cache then what will happen? > Shouldn't squid start deleting those cached objects which are not in > use? > Which parameter define that? cache_swap_low and cache_swap_high define, when will squid start removing objects and how much will it remove. cache_replacement_policy defines, how will squid choose which objects to remove. Note that for 36GB disks I have no problems defining cache_dir 30000 (the FS becomes filled up to ~86% which is OK) > What parameters do I need to check to get best performance and caching. increase cache size for bigger efficiency, but not too big (filesystems filled up too much use to have low performance). > My squid is currently handling 80 req/sec and during peak hours nearly > 1500 users uses caching server. > I can save maximum 15 % to 18 % bandwidth. Is it possible to save more > bandwidth with better performance?? yes, do the above, choose good replacement policy (heap LFUDA or heap GSDF) and increase maximum_object_size (for 2x36 GB I'd use 32 MB and maybe more). -- Matus UHLAR - fantomas, uhlar@xxxxxxxxxxx ; http://www.fantomas.sk/ Warning: I wish NOT to receive e-mail advertising to this address. Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu. If Barbie is so popular, why do you have to buy her friends?
