[c-nsp] Dual Sups and failing

[Brad Swanson]

Consider the following set up:

data center, standard Cisco "U" shaped design...so no spanning tree
Dual supervisors in 6509 (either sup1a and sup720-3a, both setups had this 
happen)
cat os - sup1a running 6.4.4a and the 720 is runiing 8.2.2
using only L2
set sys high enab is used, works fine when one sup really does die (sarchasm 
perhaps)

The issue is we have seen in both a sup1a dual set up and sup720 dual set up 
that a sup doesn't really just die, it tries to hard to hang on.  In the 
sup1a case, it was the primary sup that tried to constantly reset itself via 
software, and then the standby did in fact take over but we kept getting 
SYS-0-MOD_NOSLCPRESPONSE:Module 1 SLCP not responding... resetting module 
errors and this kept dropping mod 2 off as well...so the swt would drop off 
the network.

In the sup720 case, the primary in slot 5 started barking about an inband 
error from mod 6, the standby, so then mod 5 would rrun a self test that 
again, drops it off the network while the test is being run.  Over and over 
and over again...

So, is there some Cisco cmd that we can use that looks at a dieing blade and 
after 2-3 module resets, the primary sup can simply drop power to the 
not-quite-dead sup to FORCE it down hard?  Is this possible now or perhaps 
in some upcoming code releases?  Does native mode provide some better 
functions?

Thanks,
brad swanson