[c-nsp] MST operation...

Steve Fischer sfischer1967 at gmail.com
Mon Mar 31 22:57:46 EDT 2008


I am running (2) Cat6509-E's with Sup720-3B's running IOS.  They are
connected via layer 2 by a (2)10GigE port-channel. Spanning tree is
configured via MST with 3 instances - instance 0 (default), instance 1
(roots all odd-numbered VLANs to switch 1 - priority 4096), and instance 2
(roots all even-numbered VLANs to switch 2) - pretty simple configuration.
Switch 2 is the secondary for odd-numbered VLANs (priority 8192), and the
same is true for switch 1 on the even-numbered VLANs

All was well, but we recently upgraded the code from 12.2(18)SXF12a to
12.2(18)SXF13 to address vulnerabilities Cisco published - not a quantum
leap in terms of code revision.

 

Now, the root of MST0 is properly situated, but both switches think they are
the root for MST1 and MST2.  I cannot, as yet, link this change in the
operation of spanning-tree to the code upgrade - this is in a lab scenario
for the time being.  Debugging of spanning-tree events, root, and bpdu's
revealed nothing occurring across the port-channel.  The operation of the
Port-channel seems to be fine from all reports on the switch.  Even had a
couple of CCIE's at the VAR look at it, and nothing jumped out at them as to
being obvious.  The switches were rebooted a couple times, and the MST
configuration was cleared, and re-entered into the switch. 

 

Show spanning-tree MST detail reveals that packets are being exchanged
between the two switches on MST 0 over the port-channel, but on MST's 1 & 2,
but switches show transmits, but 0 receives across the port-channel.  This
has me a bit baffled, and thought I'd throw it out to this forum to see if
anyone has seen similar behavior.

 

Any and all insight and assistance in getting to the "root" cause of this
(pun intended) is most sincerely appreciated.

 

Regards, 

 

Steve Fischer 



More information about the cisco-nsp mailing list