you have a rather huge percentage of i/o wait. this could be caused by 
slow disks writing the kannel log, or cheap rtl81** nics doing polling.
try to use faster disks, or intel / 3com and so on nics.

On Thu, Dec 08, 2005 at 02:09:38PM +0200, Deon van der Merwe wrote:
> Hi,
> 
> We have been running a kannel service for about a week now.  The setup is:
> - single server
>  + 2Gb RAM
>  + Intel(R) Xeon(TM) CPU 3.20GHz
>  + SCSI RAID1
> - 2 kannels
>  + each of them has a bearerbox and a wapbox
>  + each kannel instance gets different users based on the users IP address
>  + both kannels connects to a single local squid proxy
> - Standard FC4 installation
> - Doing about 29tps (sent) at peak times
> - Doing about 20tps (received) at peak times
> 
> If you take a look at the attached graphs, you will see the message
> rates that it is running at.  The graphs are generated from the admin
> interface.
> http://www.truteq.com/download/cpu-day.png
> http://www.truteq.com/download/KannelStatus1-day.png
> http://www.truteq.com/download/KannelStatus2-day.png
> http://www.truteq.com/download/RadiusSessionCount-day.png
> 
> My question is this: how does the load (normal Linux load*100) compare
> to systems that you are running?
> Do you have any tips for reducing the load?
> Should we be worried about the load as it stands now?
> 
> Here is also a snapshot from top:
> top - 13:17:09 up 37 days,  1:08,  8 users,  load average: 0.94, 0.71, 0.60
> Tasks: 119 total,   1 running, 118 sleeping,   0 stopped,   0 zombie
> Cpu(s):  5.8% us,  1.5% sy,  0.0% ni, 83.6% id,  8.6% wa,  0.2% hi,  0.3% si
> Mem:   2074824k total,  2023448k used,    51376k free,    79768k buffers
> Swap:  2096440k total,     1540k used,  2094900k free,   989360k cached
> 
>   PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
>  1518 kannel    15   0  527m 410m 1956 S  4.7 20.2  35:56.27
> /usr/local/sbin/wapbox -v 4 /home/kannel/conf/kannel.conf
> 13527 kannel    15   0  301m 186m 1828 S  4.0  9.2  33:03.74
> /usr/local/sbin/wapbox -v 4 /home/kannel/conf/postpaid.conf
> 24110 squid     15   0 38960  35m 1828 S  3.0  1.7 138:05.56 (squid) -D
>  1532 kannel    16   0  128m 3420 1156 S  1.3  0.2  17:57.43
> /usr/local/sbin/bearerbox -v 4 /home/kannel/conf/postpaid.conf
>  1505 kannel    16   0  128m 3220 1148 S  1.0  0.2   6:56.03
> /usr/local/sbin/bearerbox -v 4 /home/kannel/conf/kannel.conf
> 24145 root      16   0  2024 1024  788 R  0.7  0.0   0:00.06 top
>     1 root      16   0  1748  572  492 S  0.0  0.0   0:03.59 init [5]
> 

Reply via email to