One nic hangs whole switch? normal or model dependant?

Recently i noticed a whole switching crashing because what appeared to be a malfunctioning NIC.

Is it normal the whole switch locks up? And not only the port to which the NIC is connected?

I'm having this with the enterprise models of 3Com. And seeing a whole network going down for one bad NIC is very irritating. I'm not even talking about identifying the NIC .

I'm planning for CISCO switches in the future. Do they have the same problems?

Do all switches have this?

If not, is it possible to check this in the product properties before buying and does anyone knows the exact name/code to look for?

Thx,
GrMbL

When you say the 'whole switch locks up' I suspect you're describing what I call the 'Christmas tree', where every light comes on and flickers so fast you worry the whole thing is about to explode. If so, what you're seeing is a broadcast storm because switches by nature segment normal traffic but propogate broadcasts. Many switches have port-based settings which let you throttle the maximum broadcast rate and as long as you set the threshold sensibly it can be a big help in preventing this. I would check the manual for the switches you've got to see if there is such a feature. Decent Cisco models certainly provide it

setup is:

router--switch1--switch2
the NIC which causes the problem is connected to switch2.

Switch2 has broadcast storm control and threshold limits set at 32000. I know I wasn't clear before but shouldn't broadcast storms disrupt all switches/whole network?
All connectivity is lost on switch2 when the NIC is connected. And is resumed when NIC connects. the Christmas tree effect you describe is only for the one port with problems, that's how i noticed it in the end.

Maybe that NIC is causing a broadcast storm, switch handles it correctly because no other ports are showing the same effects, but it stills locks up, maybe because of a internal problem?

EDIT: (Latest firmware didn't solve it=NOT TRUE) EDIT
I installed latest firmware now and enabled flow control, which should maybe prevent this problem. Will check with "bad NIC" when i've time and report back)

Thx for the fast reply,
GrMbL

router--switch1--switch2
the NIC which causes the problem is connected to switch2.

Let me just have a clear idea, This NIC is a port on switch2 it self? or is it a host's NIC connected to switch2?, or is it a port on switch1 connected to switch2?

it's a host's NIC connected to switch2

Then I think there is a very good probability that it's a broadcast storm as TheBishop suggested. Otherwise, the host could be sending traffic to random MAC addresses continuously (in this case it will be broadcasted too), or it could be sending giant frames. In all cases, The cpu and/or ports of switch2 will be kept extremely busy that other traffic will have minor chances of passing.

I'd check the host for viruses or possibly hanging processes. If you suspect there is a hardware problem with the NIC, go to the host's BIOS and sit there for a while, see if switch2 still hangs, if it does then the problem is surely hardware since there are no programs sending/receiving while on BIOS.

By the way, what happens when you connect that NIC to switch1?