2950T Uplink CRC Errors

[daniel.kline]

I swapped out a 2950T-24 switch on Friday evening because it was deemed defective by TAC. All seemed well that evening, but when the users came in Monday morning and began exercising the network, I began to experience serious issues.

There are three switches (switch1, switch2, and switch3). Switch1 is the root bridge and interconnects the other two switches, and also connects the servers and firewall.

The uplink ports that connect Switch2 and Switch3 to Switch1 flash amber intermittantly. At the same time the CRC errors are increasing on these uplink ports. I'm assuming that the ports are resetting because of the large number of CRC errors.

I suspect there may be a cabling issue somewhere on the network, but I’m not quite sure where. But I want to ensure that the switches are not causing the CRC errors before I have all the cabling certified.

I have to say that this network was running on some older 3Com switches and was locking up every couple of weeks - which is why we swapped out those switches. The Cisco switches, however, are locking up evey 10 or 15 minutes. So we had to fall back to the 3Coms.

I tried moving and replacing uplink cables and reviewing interface statistics on the Cisco switches to see if could identify the source of the CRC errors or isolate the problem. But I am only seeing the input (CRC) errors on the uplinked ports of switch2 and 3. Switch1's uplink ports and the user ports on all three switches show no errors.

I have attached the show tech output from all three switches for your reference. Switch1 is the root bridge, which is where all of the servers and the firewall are connected. Switch1 is also the RMA switch sent to me by Cisco. Switch2 and Switch3 appear to have older code. I am wondering if Switch 2 and 3 are suffering from the same or similar defect as switch1?

Any suggestions would be greatly appreciated.

[Georg Pauwen]

Hello Daniel,

make sure that the ports connecting the switches have the same duplex and speed settings. If possible, configure the ports for fixed settings (no autonegotiation).

I would start with replacing only the uplink cables and see if that makes a difference.

You can also issue the ´show post´ command to check for any errors during the boot sequence of the switch.

Also, and this might not apply to you, since you have already RMAed the switch, there is an issue with certain 2950 switches. Can you check your switch log for ´SCHAN ERROR INTR´ errors ? If those errors are occurring, you might have encountered the following bug:

CSCdv83336 Bug Details

Under certain level of traffic load, the (2940) switch will start logging the following messages on the console:

SCHAN ERROR INTR: SRC=6 DST=5 OPCODE=20 ERRCODE=5

and after a few seconds, the switch will stop passing any traffic. In some

cases, the switch seemed still forwarding broadcast and multicast traffic,

which will cause STP problem if the switch has redundant link and is not

supposed to be the root for the VLAN, as both port will go forwarding.

The same error message has been identified in CSCdu87836.

An assessment of the impact

Unit stops passing any traffic.

WORKAROUND

Several units were returned by CISCO. The units were re-screened to the

latest test program, and failed the SDRAM memory test.

Customer should RMA unit back to Cisco.

Regards,

GP

[glen.grant]

First i would clean up your uplink settings, some have hardcoded speed and duplex , others you only have hardcoded the speed . Also check your spanning tree root and make sure it is where you think it is , just because you put a priority statement in one of the switches doesn't mean it became root , I would advise putting spanning tree vlan 1 root , this will put the priority at 8192 and it is highly unlikely anyone else would then become root . Also look for loops in the network . Also make sure all clients are set for portfast . This really sounds like some sort of wiring problem if you were having problems on the 3coms also . I don't know how far away from each other the switches are but maybe you should look for length issues on the connecting links . The weird thing I see is switch 1 and switch 3 appear to think they are root for vlan 1 , switch 2 doesn't know who is the root for vlan 1 .

[daniel.kline]

The uplinks on switch 2 and 3 only appear take the speed command, not the duplex command. I thought the same thing about hardcoding these settings, but the Cisco documentation indicates that the 1000BasT ports only use full duplex.

I discovered that switch3 is experiencing SCHAN ERRORS - this is an RMA condition. This may be why Switch3 thinks it is also a root bridge. I am trying to determine if the third switch also has this defect.

Regards,

dk

[sebastian.wiegand]

Hi Daniel,

Would u please explain what a Media Type of T is?

Regards,

Sebastian