Rapid activity on all switch ports...

mattipler

I’ve 5 Dell (urghhhhhhhhhhhhhhhhhh) power connect switches. All up linked with fiber. Two of the switches are experiencing rapid LED activity on all ports… these two switches are directly connected and the LED activity lights on both are pretty much flickering solidly! The switches are VLANed with 100mg phones and 1gig pc ports. However this doesn’t seem to be propagated to the other switches? Incorrect MAC-address table entry forwarding packets between the two switches perhaps? Reckon I should clear the MAC addresses? I’ve tried bring all the switches offline one at a time then back online one by one but with no joy. Any comments would be greatly welcomed.

miller811

Typically that is seen when a loop exists.

mattipler

We've spanning tree active on the switches...

gojericho0

It could be a broadcast storm with one of the connected devices. Also was any new wiring or patch cables plugged in. If you have a bad one it could be causing some havoc as well.

Can you remote into the switch or is it completely pegged?

seuss_ssues

Put the switch into managed mode, remote in, and check the logs. By checking the logs you can check to see if you have a large amount of broadcast or multicast traffic.

networker050184

Your best bet would be to switch them out with some Cisco gear!

mattipler

There's a lot of collisions on one of the two switches... In the millions on a few ports that service our direct connection to the internet.

What confuses me more than anything is that this is happening on all ports of both switches... regardless of what VLAN the port is a member of!!!

gojericho0

Can you display a sh int on one of these interfaces? Are they running in half-duplex?

mattipler

All ports running in full... as they are on all the switches, including those without the rapid port activity.

The two switches that have the rapid LED activity still have the rapid activity on the uplink LED's when all devices are disconnected. I'm wondering whether it has something to with the software version these two switches are on (which is different to our other switches but the same as one another). Perhaps this could be causing the rapid activity? I've ruled anything device related. I've ruled out loops. I've ruled out MAC table inconsistancies.

Been in work looking at this for 11 hours now and feel like i'm hitting my head against a wall.

Thanks for your help thus far guys...

networker050184

Are you sure there is rapid activity or is the light just blinking rapidly?

I looked in the Dell documentation and there is nothing about difference in activity LED speed. It could just blink rapid all the time with the software version...

boostinbadger

I would analyze it with Wireshark. That would be able to nail down your problem real quick. I battled one of these recently at a jail I service. It was a broadcast storm caused by a HP Procurve. When they added on and put in new fiber they didn't disconnect their old tranceivers which caused a switching loop. Needless to say the HP was not configured with ST. I hooked up Wireshark and saw exactly what the problem was.

tech-airman

mattipler wrote:

I’ve 5 Dell (urghhhhhhhhhhhhhhhhhh) power connect switches. All up linked with fiber. Two of the switches are experiencing rapid LED activity on all ports… these two switches are directly connected and the LED activity lights on both are pretty much flickering solidly! The switches are VLANed with 100mg phones and 1gig pc ports. However this doesn’t seem to be propagated to the other switches? Incorrect MAC-address table entry forwarding packets between the two switches perhaps? Reckon I should clear the MAC addresses? I’ve tried bring all the switches offline one at a time then back online one by one but with no joy. Any comments would be greatly welcomed.

mattipler,

Questions:

1. Which model switches are these?
2. Let's call them SwitchA through SwitchE. Exactly what do you mean by "all up linked with fiber?" I'd like to know a list of [SwitchX][port#]---[cable type]---[port#][SwitchX]
3. Are all 5 switches involved in a "router on a stick" configuration? If not, how are the switches related to each other? How is routing provided to the switches?
4. When you mention "Two of the switches are experiencing rapid LED activity on all ports...", using the convention of SwitchA->SwitchE, which two specific switches are involved?
5. What are the VLAN maps of VLAN to switch/port relationship?
6. When you mention "....However this doesn't seem to be propagated to the other switches..." are there any VLANs on the "...Two of the switches..." that are NOT on the "....other switches...?"

malcybood

Not that familiar with Dell switches, but it does sound to me like a loop of some sort and potentially the STP configuration.

How is STP configured?

Have you activated RSTP (802.1w) on the uplinks that constantly blink? When you say direct connection to the internet, what type of device are these ports connected to..... router?

portfast on cisco switches should only be enabled on access links to a single device, otherwise it can cause a loop as it goes straight to forwarding mode not sure about Dell.

If you enable portfast / RSTP on an Cisco switch port, you get a warning message advising not to set the port to portfast unless a single device is attached. If the port is connected to "hubs, switches, concentrators, bridges etc.. to this interface when portfast is enabled can cause temporary bridging loops....Use with CAUTION"

Might be worth checking on the Dell.

I have also read a few web pages about STP being brutal on Dell switches interoperability wise.

I guess a simple test to find if STP is related is to disable it on these interfaces? If the problem goes away, then it lies with STP config or interoperability, if not then you explore another option.

Remember to enable it again after the outcome of the test....incase a user decides to plug a cable from one wall point to another, been there before - not nice!!!

malcybood

.....and here's the dell documentation that says pretty much the same thing as Cisco

http://supportapj.dell.com/support/edocs/network/PC5316M/en/CLI/spanning.htm#1013521

try doing the below specifically on the 2 problematic interfaces and any other uplink interfaces that are connected to anything other than a PC or "standard user" port

console(config-if)# no spanning-tree portfast