Re: OT: Cisco VSS

From: Rick Mur <rmur_at_ipexpert.com>
Date: Sun, 22 Nov 2009 15:55:11 +0100

Hi,

I also did quite some implementations and have another on the roadmap.
I'm having overall good experiences with the technology. Still the older releases especially the modular ones could be unstable. I'm having great experience with the Modular SXI2 release. An important memory leak bug is solved in that one, which caused a major customer network to just die :-)

But indeed, you need to change your design standards, it's something totally different. You also need to make sure that every device connected to your VSS cluster with at least 2 cables, otherwise the concept doesn't work. I also think it positions better in the distribution layer than the core sometimes.

-- 
Regards,
Rick Mur
CCIE2 #21946 (R&S / Service Provider)
Sr. Support Engineer  IPexpert, Inc.
URL: http://www.IPexpert.com
On 22 nov 2009, at 12:31, Travis Niedens wrote:
> I've deployed a few VSS cores.
> 
> 
> 
> 1. make sure you get the proper hardware for the VSS.
> 
> 2. Make sure to have at least 2 VSLs (sup/sup, 10G mod/10G mod).
> 
> 3. Be careful with code upgrades. SXH 1 - 3 was deferred due to nasty route
> update (or lack there of) issues.
> 
> 4. Keep in mind older line cards may not work with VSS.
> 
> 5. Once the VSLs are up and the VSS is active, DO NOT CHANGE THE VSL PORTS.
> The config may seem wrong but it is not.
> 
> 6. Up until recent SXI code versions, the only ways to detect a "split brain"
> scenario were PaGP+ dual active and BFD.  To me, these approaches are like
> cutting a leg off to help someone walk again -- doesn't really work and is
> prone to issues. A bug / issues with a far end switch running PaGP+ could
> easily impact a VSS core as the response to shut down one WHOLE core (sans
> ports specified no to). The clean up to bring the core back to full capacity
> can be longer than just rebooting the whole core (I have seen this on 2
> occasions).
> 
> 7. If you run latest SXI, run fast hellos for VSS dual active detection.
> 
> 8. Make sure your services modules are running the latest code that is stable
> and supports VSS.  VSS support for services modules came in SXI.
> 
> 9. Adapt your design model to handle the changes in VSS (no HSRP, no spanning
> tree primary / secondary root, etc.) Also, single attached services can have
> issues during a dual-active event where a core is shut.
> 
> 
> 
> Hope this helps !
> 
> 
> 
> Travis
> 
>> Date: Sat, 21 Nov 2009 08:49:25 -0600
>> Subject: OT: Cisco VSS
>> From: ccie.09_at_gmail.com
>> To: ccielab_at_groupstudy.com
>> 
>> I am planning to upgrade two Cat6513 to VSS (inter chassis link through
>> Sup720). Any one with experience can share some pros and cons please?
>> 
>> 
>> Blogs and organic groups at http://www.ccie.net
>> 
>> _______________________________________________________________________
>> Subscription information may be found at:
>> http://www.groupstudy.com/list/CCIELab.html
>> 
>> 
>> 
>> 
>> 
>> 
>> 
> 
> _________________________________________________________________
> Windows 7: It works the way you want. Learn more.
> http://www.microsoft.com/Windows/windows-7/default.aspx?ocid=PID24727::T:WLMT
> AGL:ON:WL:en-US:WWL_WIN_evergreen:112009v2
> 
> 
> Blogs and organic groups at http://www.ccie.net
> 
> _______________________________________________________________________
> Subscription information may be found at: 
> http://www.groupstudy.com/list/CCIELab.html
Blogs and organic groups at http://www.ccie.net
Received on Sun Nov 22 2009 - 15:55:11 ART

This archive was generated by hypermail 2.2.0 : Tue Dec 01 2009 - 06:36:29 ART