Baystack 450 cascade issues 1

[cityofshawneeok]

I have a stack of three switches (450-24T) in a closet that is suddenly having cascade issues. The base unit has a 400-2fx MDA module.

Here's the log from one of the units that seems to have started the issue. This unit is no longer in the stack but problems persist.

-----------------------------------------------------
Entry Number: 6 sysUpTime: 00:05:16 Reset Count: 46
Down (In) Cascade Cable Disconnect Detected

Entry Number: 5 sysUpTime: 123 days, 06:56:11 Reset Count: 44
An operational exception occurred at vector number 2.
[SP+0] 2004000B [SP+3] 0000FFFF
[SP+1] 6AE6C008 [SP+4] 000B6AE6
[SP+2] 00C00000 [SP+5] 00010055

Entry Number: 4 sysUpTime: 123 days, 06:55:19 Reset Count: 44
Down (In) Cascade Cable Disconnect Detected
-----------------------------------------------------
The other units only reported the cascade disconnect and when looking at the down stack the cascade LEDs are all green. Since that initial 'cascade cable disconnect' and the exception, the stack continues to fail with a blinking base light between 20 to 45 minutes or more after re-powering. Before upgrading the OS I would get the cascade cable discon errors and the stack would fail with only the base unit passing traffic to the fiber MDA.

I have since reset each switch to factory defaults and upgraded the switches independently to 4.5.2.4, have replaced each switch with a backup unit, swapped cascades modules and cables but can’t pin down the error. For now I’m running crossover cables to bypass the cascade until I either resolve or replace the whole stack.

Anyhow have something similar? Just trying to find what I might have overlooked.

I’m just about ready to throw the stack out and buy a complete replacement 5xxx setup.
Thanks
-Stephen

 

[netmanrick]

Pull out the cascade adapter from th 450 that's not in the stack.Power reset the unit.Have a pc on the console port to watch the selftest messages.
Do you have the cascade modules installed in the units with xover cables ?
I am ASSUMING you verified the base switch setting on the cascade modules.

Rick Harris
SC Dept of Motor Vehicles
Network Operations

 

[cityofshawneeok]

Thanks for the reply.

Yes, the base switch settings are good and have been double checked. The stack comes up without issues, identifies the base and works at first but then has the blinking stack configuration error later.

I've watched each switch to insure it selftests properly and none report problems.

Actually yes right now I have the xover cable in and have left the cascades in since I did not want to re-power the switch in the middle of an active day and it doesn't think it's a stack anymore anyhow.

I've run it without the cascades with xover cables last week when it failed until I could reconfigure. It has to be some type of failure since the stack worked fine for a couple years without issues but I can't track it down even with the swaps I've done.

I wish I knew more of what that "vector 2" error was. That switch is no longer in the stack and I still have problems so I'm wondering if it created other issues with the switches that were in the stack with it and are still in service.

-Stephen

 

[netmanrick]

Are the switches all at the same fw & sw rev ?
Last resort, init each sw to defaults & reconfigure.

Rick Harris
SC Dept of Motor Vehicles
Network Operations

 

[cityofshawneeok]

Yes, all are in sync, have now been reset and upgraded to latest OS as well. The stack ran as configured for years so obviously something else has failed beyond the one swtich that has been replaced but I cannot seem to find the issue. I appreciate all comments. I think I am going to replace this stack and then bench the the existing stack in a non-critical situation and see how it performs.

-Stephen

 

[norteldude78]

FYI: A vector 2 error is a software error. The numbers it gives after it are protocol numbers and memory locations. A software upgrade usually solves it.