[c-nsp] ASR920 is a ticking timebomb (CSCvk35460)

[James Jun]

> 
> Has anyone been hit by CSCvk35460?

We have definitely been hit by this bug on two ASR920s running 03.16.02aS with over 2 years of uptime.

The symptoms and effect of the bug from our observation include:

- packets continue to forward, so there is no immediate disruption to traffic
- interface counters stop increasing, snmp probing of the box fails with all interfaces returning 0
- interface link state stops.  if an interface goes down (e.g. due to fiber cut or someone unplugging link), port will continue to remain up on IOS XE.  this creates lot of annoying operational problems.
- due to the above, new interface being plugged in will not come up, so you can no longer provision new ports.
- DOM monitoring dies.  attempting to show transceiver stats will freeze the terminal for several seconds then error out.


Our reaction to this bug was to simply upgrade SW to at least 03.16.04 or newer (03.16.06 and .07 seem to have completely broken MPLS FRR) and reload, but it seems this bug is not fixed at all?

James