request per second per cache is around 18 at peak time. traffic flow combined to all three cache is around 5.7Mbits at peak. do you think this cpu usage is normal for the provided load ?? On Mon, 27 Nov 2006, Adrian Chadd wrote: > On Mon, Nov 27, 2006, Manoj Rajkarnikar wrote: > > Hi, > > > > I'm seeing some unusual cpu usage pattern in our cache farm as well. we > > have 3 cache connected to one wccpv1 router: > > Whats the request rate at this point in time for each of the Squids? > Check the 5min counters in Cache Manager for an idea and let me know. > > You've given me three caches here which have different CPUs and different > architectures; of course they're going to have different CPU usage patterns. > CPU usage is a combination of raw Squid processing, along with network traffic, > disk usage, disk and network drivers (the amount of time spent in interrupt > and IOWAIT here gets factored into the loadav calculation but not always > in the process % usage!), etc. Its quite complicated! > > Don't forget it also takes CPU to recycle the disk and memory caches; if you've > not given it enough disk space then you might find a lot of IO being wasted > on clearing that. > > > > > > Adrian > > > 1. 2.66GHz P4 with 1GB RAM, 128M squid mem, 12GB aufs cachedir. RHL 7.3 > > > > 10:57am up 117 days, 2:26, 1 user, load average: 0.35, 0.35, 0.31 > > 65 processes: 64 sleeping, 1 running, 0 zombie, 0 stopped > > CPU states: 0.2% user, 0.1% system, 0.0% nice, 0.1% idle > > Mem: 1015748K av, 1004084K used, 11664K free, 0K shrd, 12368K > > buff > > Swap: 1052248K av, 632K used, 1051616K free 32908K > > cached > > > > PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND > > 803 nobody 17 0 326M 326M 1320 S 55.1 32.8 33841m squid > > > > 2. 3.2GHz P4 HT with 1GB RAM, 128M squid mem, 12GB aufs cachedir. RHL 7.3 > > > > 10:57am up 116 days, 14:23, 1 user, load average: 0.91, 0.79, 0.44 > > 63 processes: 62 sleeping, 1 running, 0 zombie, 0 stopped > > CPU0 states: 3.1% user, 18.0% system, 0.0% nice, 77.1% idle > > CPU1 states: 0.1% user, 11.1% system, 0.0% nice, 87.0% idle > > Mem: 1023688K av, 1011440K used, 12248K free, 0K shrd, 21424K > > buff > > Swap: 1052248K av, 19808K used, 1032440K free 27100K > > cached > > > > PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM TIME COMMAND > > 24046 nobody 16 0 287M 284M 14120 S 18.9 28.5 3785m squid > > > > 3. 2.8GHz P4 with 1GB RAM, 128M squid mem 12GB aufs cachedir. CentOS 4.4 > > > > top - 10:53:56 up 46 days, 31 min, 3 users, load average: 0.07, 0.06, > > 0.09 > > Tasks: 49 total, 2 running, 47 sleeping, 0 stopped, 0 zombie > > Cpu(s): 2.0% us, 1.0% sy, 0.0% ni, 92.0% id, 4.3% wa, 0.0% hi, 0.7% > > si > > Mem: 1034528k total, 1021380k used, 13148k free, 126964k buffers > > Swap: 2096472k total, 112k used, 2096360k free, 430644k cached > > > > PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND > > 5152 nobody 15 0 374m 365m 1424 S 3.3 36.2 676:22.76 squid > > > > > > All of the above squid boxes receive equal amount of hash bits from wccp. > > can you link anything ?? Thanks. > > > > > > On Mon, 27 Nov 2006, Adrian Chadd wrote: > > > > > On Sun, Nov 26, 2006, Wojciech Puchar wrote: > > > > getting this under moderate load (highest is 2* this). > > > > > > > > 2/3 of time is system time. is it OK? can squid be made faster. it's using > > > > diskd, disks are much less loaded than CPU > > > > > > Which version of Squid are you using? Whats the request rate at the time > > > Squid is using that much CPU? Whats the hardware spec? > > > > > > > > > Adrian > > > > > > > > > > -- > > Manoj Rajkarnikar > > > > System Administrator > > Vianet Communications Pvt Ltd > > Pulchowk, Lalitpur, Nepal. > > (PH)977-1-5546410 > > -- Manoj Rajkarnikar System Administrator Vianet Communications Pvt Ltd Pulchowk, Lalitpur, Nepal. (PH)977-1-5546410
