procurve ip route on flat network

[Stuart Brand]

We have a major problem on our network, the core switch looks like it is flooding packets, I had no involvement setting up the core but I have been asked to look into this problem, the people who set it up have also left

the main thing that strikes me is that ip route is enable and static routes in place on a flat network? see below

could this be whats causing the storm?

ip default-gateway 172.16.3.1
ip routing
logging 172.16.16.11
snmp-server community "public" Unrestricted
vlan 1
name "DEFAULT_VLAN"
untagged A1-A22,B1-B22,C1-C22,D1-D6,E1-E6,F1-F6,G1-G6,H1-H6
ip address 172.16.6.1 255.255.0.0
exit
interface A18
monitor
exit
fault-finder bad-driver sensitivity high
fault-finder bad-transceiver sensitivity high
fault-finder bad-cable sensitivity high
fault-finder too-long-cable sensitivity high
fault-finder over-bandwidth sensitivity high
fault-finder broadcast-storm sensitivity high
fault-finder loss-of-link sensitivity high
ip route 172.16.96.0 255.255.240.0 172.16.1.4
ip route 172.16.80.0 255.255.240.0 172.16.1.6
ip route 172.16.64.0 255.255.240.0 172.16.1.3
ip route 172.16.112.0 255.255.240.0 172.16.1.5
no stack
spanning-tree
spanning-tree priority 0

[Matt Hobbs]

I would be looking at spanning-tree first, verify the topology is stable by checking the 'time since last topology change' from the 'show span' command.

If the spanning-tree is not stable, I would update the firmware on all the ProCurve switches in this network. Actually I would do this regardless.

The static routes in the same VLAN are not ideal, but shouldn't cause the problem you describe. It will cause a fair few ICMP redirects though when a host is trying to connect to one of these other networks. I would look at creating separate physical VLANs for these other routers later on.

If the network is really in bad shape at the moment, you may want to start troubleshooting it the old fashioned way - start removing cables one a time until the problem is gone, and then follow the problematic cable(s).

Finally, is this a 4100 switch? (I can see the 'no stack' command and from memory it's the only modular switch that supported stacking). If so, it simply may be being pushed too hard - it's not the most powerful router out there. How is the CPU utilisation on this switch?

[Stuart Brand]

Hi

I think I've found the main cause, the core only seems to be holding about 500 mac address? I think it has a capacity of 8000, how would I check that?

Thanks

[Ardon]

What IS your Core Switch? Can not be the HP one. One VLAN and IP Routing enabled why?
Seems like a flat network to me.
Also you need to elaborate on the "the core switch looks like it is flooding packets" statement. What do you see?

I would focus more on your Core Switch then the output of the switch above.

Topology and Show Tech All files would be very handy.

[Stuart Brand]

ok, we have a core switch which is a HP 4108GL and it is a bad design, all the servers are connected directly, and there is no distribution layer :(

Yes this is a flat network and for some reason they added static routes to the core even tho there is no vlans and no other routes off the core

I have narrowed down the problem, it is only windows 2000 and NT 4 clients that effect the problem, the core is ignoring the mac address and any packets destined of those clients are flooded out of every port on the core,

the core cannot ping those clients either, it just isn't picking up the arp information :(

the firmware version is .93 and the latest is .103 so it is ten versions out of date, however looking at the release notes there are no address's to this problem so my manager is reluctant to update the firmware incase something goes wrong and we don't currently have a replacement core switch :(