On 1/08/2013 9:06 p.m., Ahmad wrote:
hi , i have another question here , with my config above with cache_mem is 1000 , now if i chance cache_dir to 150 G instaed of 90 G for the 3 hardsisks above , which result in total about 150* 3 Giga storage . ===> i note that squid wccp service gett flapped every about 40-50 minutes and seems squid to be unstable . but why it affected while im fixing cache_mem to 1000 ??????? is my thought correct ??????? i mean that im limiting squid_mem to 1 Giga , but i need about memory usage of 150*3*15 = about 7 G shouldn't squid be stable ?? or it is normal result ??
The WCCP flapping is caused by Squid being too busy with something to send out its 10s repeated HERE_I_AM update to the router for a few in a row. If that is triggered by simpy setting large disks it is likely the CPU cycles consumed by disk I/O and/or garbage collection taking up Squid's time.
Amos
