Date: Mon, 9 Feb 2004 22:15:35 +0545 From: "Bikrant Neupane" <bikrant_ml@wlink.com.np> To: <squid-users@squid-cache.org>, <freebsd-isp@freebsd.org> Subject: High Cpu utilization Message-ID: <006001c3ef2a$0e029120$0100a8c0@ekanta4ey6dw38>
next in thread | raw e-mail | index | archive | help
Hi, For some reason my squid cache is running at 99% load. I don't see any abnormal pattern in traffic and number of connections. Agv. traffic is around 2M/sec and goes upto 3.5M/sec during peak hours. Max File descriptor open is around 2K. I'm running squid-2.5.stable4 on FreeBSD-4.9. The hardware is Dell Power edge with 2.6 Gig P4 HT processor and 2 GB RAM. I've got custom kernel running with HT enabled. I've got 2 cache dirs (10 Gb each) on 20 GB SCSI drive dedicated for cache only. This system has been running since last 30 days. Under same traffic load (and connections) the avg. cpu usage was around 30% during peak hours. but since last 2 days it is running at 99% most of the time :( I compiled squid with following options: ./configure \ --enable-delay-pools \ --enable-ipf-transparent \ --enable-storeio=diskd,ufs,aufs \ --disable-ident-lookups \ --enable-snmp \ --enable-wccp \ --enable-removal-policies \ -sysconfdir=/etc/squid/ \ --with-pthreads \ --enable-async-io=60 \ --enable-async-io \ --enable-time-hack And these are few lines from my squid.conf: cache_mem 128 MB cache_swap_low 90 cache_swap_high 95 maximum_object_size 32768 KB maximum_object_size_in_memory 16 KB tcp_recv_bufsize 65535 bytes ipcache_size 4096 fqdncache_size 4096 cache_dir aufs /usr/local/squid/cache1 8000 32 512 cache_dir aufs /usr/local/squid/cache2 8000 32 512 I think I need to tweak some parameter..Please help. with regards, Bikrant
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?006001c3ef2a$0e029120$0100a8c0>