On 25/10/17 01:34, Vieri wrote:
Hi, During the last 3 days my Squid process has gone wild with 100% CPU usage once a day. Since this is a production system (never had this issue before), I "quickly restart" the Squid daemon, and that solves the issue until the next day. Restarting takes quite a while... Next time I get a 100% squid process that brings my proxy to a crawl HTTP-wise, what can I try in order to get more info, and possibly see the cause of this? Would squidclient mgr:info shed some light on this?
That one is just the overview, so not likely to show more than that things are going slowly.
Usually the 100% CPU is an indicator that the machine is swapping. So check the memory usage when it is happening.
Next is probably the cache garbage collection. mgr:storedir report might show that. Or an strace showing a lot of cycles in the cache replacement calls.
Amos _______________________________________________ squid-users mailing list squid-users@xxxxxxxxxxxxxxxxxxxxx http://lists.squid-cache.org/listinfo/squid-users
