NLB on the webservers and server load

[Jeroen Wijnands]

I'm posting this hoping there's some new developments.

I'm looking at the load on my 2 front end webservers (I have a
standard 5 server environment with project 2007 and WSS) which are
running NLB. One of these is showing a lot more resource usage than
the other. I know NLB works well enough that when I reboot one then
the other handles the website.

Is it possible I've made a configuration error somewhere? Or is this
all I can expect from my present configuration and do I need to add
some sort of hardware load balancing device to spread the load evenly
across my two webservers?

Thanks!

 

[Gagan Goel - MCP PMP]

I have used NLBs for 2003 environemnt and it worked fine. It could be a
session affinity thing but below a helpful article which you can refer to
recheck your configuration.

1. All port rules (range) defined in a host cluster must be unique across
the cluster.
2. Host priority (Default Host) must be unique across the cluster.
3. Cluster mode must be unique across the cluster: either Unicast or
Multicast.

http://support.microsoft.com/kb/556068

Hardware could be helpful but is usually cost and resource intensive.

Gagan

 

[Jeroen Wijnands]

Thanks for trying to help Gagan,

Turns out I was worried for nothing. Had a look at the IIS logs on
both nodes, they're nearly the same size. Turns out the cpu and memory
load spikes in my monitoring were caused by a runaway snmp process
which was caused by the ibm servraid management agent.