Discussion Locked

Cisco switching issues

0 helpful
11 replies
Closed to new repliesThis discussion is closed. It stays readable, but it takes no new replies.

We have a stack of 3 cisco 2900 xl switches, all connected to a cisco fibre backbone switch. I've recently noticed when connecting printers etc that ports on two switches don't light up as a device is connected, whereas one lights up instantly. I've checked (via visual switch manager (a java based web client thing) that all ports are set to autonegotiate speeds etc and that port fast is on, but this doesn't help... any ideas???

Also , some devices are experiencing severely slow network access (i.e. copying off a 150 mb file takes 45 minutes as opposed to the usual 45 seconds).. could be related to the other problem???

11 replies

by mckaytech

It's rare, but not unheard of, to have bad switch ports on a Cisco 29xx but that seems unlikely from your description. However, you could test that theory by swapping ports and seeing if the problem improved.

For the slowness problem, I would look carefully at the switch configuration and consider using explicit speed and duplex ratings rather than auto-negotiate. In particular, you might want to verify that neither port spanning nor port monitoring are in use.

paul

by Member 29244100

If it was a 4000, 5000, or 6000 series switch, I'd suggest looking at PORTFAST, UPLINKFAST, or BACKBONEFAST values for those ports. You say you've checked PORTFAST enabled on the 2900? Is it just a PC or printer, not a hub you're connecting, so you're sure there aren't going to be any Spanning Tree Loops? This is also suggested by the slow performance? Are any of the ports going in to a 'BLOCKING' state during transfers? This could be a Fibre Uplink problem rather than a problem with the actual 10/100 Port.

I've seen a similar sort of odd behaviour before when the 2900XL is acting as a VTP Client to a 6500 VTP Server when one or either of the fibre links to the backbone switch are either not in the Gigabit Etherchannel or VTP Trunksetup on the core switch. It then takes the 2900XL and the core switch between 15 and 30 seconds longer to 'negotiate' roving VLAN memberships for the port.

Something else to check up on - that the core switch (I assume it's a 5500 or 6500) is actually the root switch for Spanning Tree.

I'd also suggest that you compare the IOS versions on the 2900XL's - are they all the same version??

Hope this helps...

Jason.

by ckilday

I know this sounds "low-tech", but have you checked to see which "mode" your switch is in? There's a small button on the left side of the machine which will put your led's into a different mode. In the STAT mode, they show normal status...on for connection and blinking for data transfer. In the UTIL mode, they react according to switch utilization and the led's come on from port 1 to port 2 and so on depending upon how heavily the switch is being used. If your switch isn't processing much traffic, only one or two lights may be on. I forget the third mode just now, but it doesn't provide the "normal" response you're looking for. Anyway, pushing the mode button will cycle it through the different settings. You can see which mode it isin as an LED will light up behind the STAT, UTIL, ? spots just above the mode switch.
As for the slowness, I agree with Mr. Wright...check the speed and duplex settings and hard code them on the port and on the NIC. Cisco does not recommend using autonegotiate due to the variances found among the NIC vendors. I'd also recommend checking your ports for errors/collisions and other symptoms of wiring/connector problems.
Best wishes
- ck

by benyap

Hi,
I am using the same type of switch in my office as well. I used to face the same problem before but after some testing, this is what we found out.

The Problem:
1. The switches is by default running in
10/100Mbps.(Autonegotiate)
2. Most network printer is running under
10Mbps network card.
The problem is the switches doesn't really autonegotiate hence the network packets sent to the printer is faster than what it can receive and this caused the lost packets.

The Solution:
Identify the ports and force the particular port to 10Mbps.

As for the slow network access, try to identify your network card if it is 10/100Mbps card. Then if you are using TCP/IP protocol, try to remove other protocol like Netware, NetBEUI and IPX/NetBIOS.

Hope this will help.

Regards

by benyap

Hi,
I am using the same type of switch in my office as well. I used to face the same problem before but after some testing, this is what we found out.

The Problem:
1. The switches is by default running in
10/100Mbps.(Autonegotiate)
2. Most network printer is running under
10Mbps network card.
The problem is the switches doesn't really autonegotiate hence the network packets sent to the printer is faster than what it can receive and this caused the lost packets.

The Solution:
Identify the ports and force the particular port to 10Mbps.

As for the slow network access, try to identify your network card if it is 10/100Mbps card. Then if you are using TCP/IP protocol, try to remove other protocol like Netware, NetBEUI and IPX/NetBIOS.

Hope this will help.

P/S : Make sure that you are using a enhanced Cat.5 cable for all devices connecting to the 2900XL switch.

Regards

This discussion is closed to new replies.