Search squid archive

eyeballed performance question

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



for those with high volume reverse proxy (http accel) experiences....
 
 I've got 8 boxes, doing a peak of about 260 req/sec each, but they are pegged at 100% disk util (iostat), which I think it legitimate, given their config.   They have two cache_dirs, each 10Gb, on two SATA disks.  cache_mem is set to 2048, it's on a 4GB 32bit linux.  aufs.  hit ratio is around 85-90%, due to being URL hash balanced.  they've got about 200k objects in hot object cache, about 2 million objects on disk.
 
 I'm about to replace them with 6-disk 15K RPM SCSI machines, same memory footprint.
 What kind of performance increase can I expect, if I were to put down a cache_dir of 10GB on each of the six disks ?  
 
 my question is more about what can I expect from going from 2disk SATA to 6disk SCSI.
 
 thanks



[Index of Archives]     [Linux Audio Users]     [Samba]     [Big List of Linux Books]     [Linux USB]     [Yosemite News]

  Powered by Linux