Broken Stack Ring

[surgebot]

Hello. 

I received the following error from SolarWinds.

The switch stack data ring on SWXX is broken. The switch stack may be up, but the data stack ring is no longer redundant and backplane bandwidth has been reduced.

Nothing has been changed on the switch.  My initial idea is to just restart both switches and see if it comes back up but my concern is that I'll lose one of them on reboot and my life will be ruined. 

When I run show switch detail I see both switches and their current state is listed as ready. I've been following what I'm reading in this  troubleshooting guide but nothing seems to geared toward a broke data ring. 

Any assistance for this new network tech would be appreciated. 

[Joseph W. Doherty]

3750s?

I recall, there's a command to show status of all the stack cable connections.  That should immediately show which connection is at fault.

A reload/reboot, likely, will not fix the issue, as the cause is often hardware related (although might be as simple as a loose connection).

BTW, besides loss of bandwidth, with one ring being "broken", if you lose another stack cable connection, the stack will partition into two stacks, which can function physically, but causes issues as the two stacks believe they are the same device.

[surgebot]

They are 3650's. 

[Reza Sharifi]

Hi,

What is the output of "sh switch stack-ring speed"?

If the second line says "full" then there is nothing wrong with switch stacking. It could be a SolarWinds software issue.

Stack Ring Configuration: Full

HTH

[surgebot · Accepted solution]

Thanks for the reply!

I ran the command.

Stack Ring Speed : 160G

Stack Ring Configuration : Full

Stack Ring Protocol : StackWise

I think I'm with you on the SolarWinds issue. I haven't had any loss of functionality or complaints about speeds from anyone and this has been going on for months. I never got around to it because nothing seemed to be affected. Now that I'm done with a bigger project I'm cycling back to the little stuff that came up along the way. 

[Reza Sharifi]

Also, removing the stack from SolarWinds and adding it back could possibly resolve the issue if it is software related.

This is probably the easiest thing to do without impacting anything.

HTH

[Leo Laohoo]

What firmware is the 3650 stack on?

Stacking with classic IOS is one thing but "stacking" with IOS-XE opens a door into a different world of nightmare.

IOS-XE and "stacking" do not mix.  There are many bugs attributed to "stackmgr" process and can only be triggered with the switches are in a stack.  If the switches are in standalone, nothing happens but the moment they are stacked you are looking at a timebomb.  

My stacks of 3850 (IOS-XE version 16.X.X) cannot withstand 6 months without crashing in between.  If my 3850 are on 3.6.X or 3.7.X, they are fine (stacked) for several years.  Once they cross the boundary of 16.X.X it is all but downhill after that.

[surgebot]

We figured it out. It's an issue with SolarWinds. Thanks for the reply
though!