Nortel 55x0 stack problem

[Klanng]

I have a stack that within 10 days have reduced itself to one unit instead of five.

The configuration is like this :

5530 (Stack master)
5510
5510
5510
5510

The stack is running 5.1.0.014 and firmware 5.0.0.4.
The stack was upgraded in the middle of january and the first crash occured february 14.
I didn't see the problem with the 'old' sw.

I seee these entries in my syslog :

Feb 23 21:19:01 :The number of stack units is 1 instead of expected 5 Info
Feb 23 21:19:02 :Leave event sent, unit mask 0x01, appl mask 0x0000000F 0xBFFFFFFF Info
Feb 23 21:18:59 :Trap: bsnStackManagerReconfiguration Info


Before I start downgrading the stack it would be nice to hear if anybody have seen this issue?

Thks

BRgds
Kenneth

 

[DaddyOfThree]

I've been running v5.1 software since it was released on a large number of standalone switches and stacks... all without issue.

Is there a cascade link failure on the stack? Is one of the UP or DOWN paths down/amber?

I would at least cold boot the entire stack before you go downgrading it.

Cheers!

 

[Klanng]

Hello,

Did a cold boot 2 days ago.
It turned out the middle switch didn't work. I had amber in switch 2 and 4.

Thanks

 

[anthonyanderberg]

Just so we're clear - did you have a return cable connecting the top and bottom switch?

 

[Klanng]

The cascade cables seems okay.
The client is currently running without the middle switch.
I have a service window this saturday where i can look into the problem. Might be a defect swicth.

 

[VmgroM]

Hi,

have you checked if the rest of 5510 on the stack (especially number 2 and 4) are not marked also as master?

I´ve been working for years with Baystack 5520, and sometimes, if you don´t review this, can have troubles, because two or more equipments try to be the master, and the stack fails.

All the best

 

[SnoProG]

Did you connect via console to the suspect switch by itself to see if any errors showed up on power up.

 

[Klanng]

VmgroM
"have you checked if the rest of 5510 on the stack (especially number 2 and 4) are not marked also as master?"

Yes - that was the first thing I checked Only 1 master.

SnoProg
"Did you connect via console to the suspect switch by itself to see if any errors showed up on power up. "

Actually for some reason I haven't tried this. I trust my syslog blindly, if it happens again this will be top of my list.

After I removed switch nr 3 - 10 days went and the problem came back. This time I did a "Reset to default setttings" and reconfigured. The stack have been running for 18 days without problems so far.

Thnks

Kenneth

 

[B0roda]

Hi,

had similar problem time ago, we did open case with Nortel.

The solution was found in defective stack cables, try to check your cables:

Cable quality defects were found in one particular cable vendor. These cables have been identified
as being manufactured and shipped from the beginning of January 2005 to July 2005.
The cable can easily be identified by the white label at one end of the cable with the Nortel part
number 216448-A, and the Vendor date code format of "Year/Day", and would be shipped in ERS
5500 products between the previously noted dates. Cables with the date code format of "Week/Year"
are not affected by this issue.

I'm from Russia...

 

[Klanng]

The cables could easily been from around January 2005 to July 2005. I will check the labels next time I pay them a visit.

Thanks for the advice.

Kenneth, Denmark ;-)