"Cisco Catalyst Blade Switch 3020 for HP" Best Practices

Unanswered Question
Apr 1st, 2009

Hi all,

Here we use this blade switch cluster for the Data Center server farm. Here we noticed some network traffic delay. when I monitor the up link I can see there is not much traffic in the link. So doubt about the traffic delay. If anyone can help me on this it would be great...

Thanks in Progress.....


  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
gnijs Wed, 04/01/2009 - 08:50


What do you mean with "you notice some network delay" ? Please explain.


chandanaa Wed, 04/01/2009 - 23:34

Here in the "blade switch for HP" a request processing application is running in a server. when we use the same type server stand alone with a separate switch it processes 80,000 requests per 4 hours. But when this blade server in the blade switch it only processes near 6,000 requests per 4 hours.

That what I mean like "Network delay"...

That is what I cannot understand.

I have attached the current configuration of the blade switch so you can go through it.

Is there any best practices specific to that type of blade switch that should enable?

Thanks for your help



gnijs Thu, 04/02/2009 - 06:09

I see you have forced the speed to 1G fixed in your configuration. You can try to remove the "speed 1000" line in the configuration and let the card autonegotiate (it will only support 1G, i know, but nevertheless autonegotiate can be important on gigabit).

For example, i have noticed that autonegotiation works very well for physical servers, however, with VMWare ESX servers we found network performance problems with autonegotiation (thoughput on copy and vmotion operations). These went away when we configured the switch interface for fixed speed. Very strange, but nevertheless...


chandanaa Thu, 04/02/2009 - 12:40

Hi gnijs,

I'll change accordingly and see....

Thanks !

I'll tell you the status




This Discussion