High CPU utilization, ping timeouts and errors in log

[goozoo]

Hello Everyone,

I have following problems. I have high CPU utilization (60-100%) on many switches and many errors like this:
"Out of pkt buffers; miss count: 1139299"
The problem concern HP AS 2000, 2524, 2512, 4000M and 8000M but not 4104GL.
Spanning Tree is enabled on all switches.
Firmware:
HP AS 2000: B.04.28
2524: F.05.59
2512: F.05.59
4000M, 8000M: C.09.28
4104GL: G.07.102
I have 20-30% of packet lost to some switches.
Ports utilization is relatively low, on average 5%.
I have trunks between backbone switches and in few places between backbone and edge switches.

Any idea?
Sorry for my English.

Regards

[Joel Belizario]

Just wondering if you have checked whether spanning tree has been reconverging or not?

[OLARU Dan]

What is different in the 4108s with respect to the other, malfunctioning, switches?

If you are sure you have no loop in the network, kill the STP. Also you should stop bells-and-whistles features such as web interface, throttling on 4000/8000, and you might want to change MAC Age time from the default 300 to 1800 seconds, disable CDP, Automatic Broadcast Control.

[goozoo]

Hello again,

In the weekend I switched off STP and CDP. I also set MAC Address age on 1800 s.
It has not helped.
I've attached configuration of my switches.

Regards,

[Matt Hobbs]

Approximately how many switch ports do you have in this VLAN? I've seen errors like this in the past when there is a lot traffic hitting the IP address of the switch itself. This can be broadcast/multicast traffic too.

If possible, I would create a separate VLAN just for switch management.

[goozoo]

I have DEFAULT_VLAN on my switches only.

[goozoo]

Thanks for all your help.
That was my fault. I have forgotten to restart one of 2524 switches after turning off STP.
Everything is working fine now.

Regards