[c-nsp] To VSS or not to VSS

[Thomas Dupas]

Hi Eric,

The FWSM (or any service module) wasn't supported in a VSS setup until SXI.
And I don't think that many people made the step yet to SXI on a production VSS system, but you never know.

Overall I have had fairly good results with VSS in terms of throughput and stability, they were mostly used as distribution switches in the campus or "bookshelf" switches in the DC. The biggest flaw so far is the downtime when performing an upgrade, you fall back from SSO to RPR due to the IOS mismatches, and that means around 5 minutes downtime on failover. Same as you would have with supervisor redundancy in a single chassis
They now have an "eFSU" (semi-ISSU?) in the SXI release, which should improve upgrade procedures (RPR+ in stead of RPR), but it's not really ISSU according to the specs. But I certainly want to try that (but then I need a next release to upgrade to :-))

Best Regards,

Thomas Dupas


On 22/11/08 21:01, "Eric Cables" <ecables at gmail.com> wrote:

I'm working on a design which includes 2 pairs of 6509s w/ VS-S720-10G
(one in each chassis).  The VSS capable supervisor engines were chosen
mainly for the 10G interfaces, but the more VSS documentation I read
the more it seems like a great solution for added
redundancy/bandwidth, while reducing complexity.  As far as modules,
all will be 6748s or 6724s, and the only service modules in the mix
will be a pair of FWSMs in one of the VSS pairs.

Can anyone provide any feedback on your VSS experiences?  How have the
FWSMs played with VSS?  Any design considerations I should be aware
of?

Thanks,

-- Eric Cables
_______________________________________________
cisco-nsp mailing list  cisco-nsp at puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/