Tyler Haske tyler.haske at gmail.com
Tue Nov 14 11:25:09 EST 2017

It isn't as grey as you'd think. I did the TAC dance for a while.

The first thing a decent engineer would do for a L1/L2 problem is pull a
"show inventory" and "show version" to verify the SW and HW combo for being
supported. This includes checking the S/N on the optics.

We would plug the S/N into the manufacturing database, it shows us when it
was built, tested, and who tested it. We also see any internal defects.

In my time half the problems were the optic itself, and half were higher up
the stack.

The problem is, Cisco can't really do anything with a 3rd party optic,
other than recommending the customer to replace it. We certainly can't find
esoteric hardware bugs that only show up when an extremely specific
bitstream pass through the data plane.

As an engineer, it's unthinkable to attempt to engage the BU for a flapping
interface with a 3rd party optic. The first thing BU escalation will do is
ask ... did you try a Cisco optic? Everyone internally will ask this,
because sometimes it's the optic!

Cases generally took two paths:

* A customer would find a supported optic for testing. This either resulted
in the case being solved outright or escalated.
* The case would get stuck until the complaining was loud enough Cisco
would borrow some optics out to the customer.


On Tue, Oct 31, 2017 at 10:49 AM, Nick Cutting <ncutting at edgetg.com> wrote:

> If you can get a third party optic to work without the command – are they
> supported by TAC?  It is a grey area for sure.

More information about the cisco-nsp mailing list