We have a pair of OS10k in VC setup (stacked over redundant 40Gig links), but we have neither took them live, nor did any big tests so far.
I'll update you in a while, after we gathered our experiences..
We had some minor *cough* problems with the VC enabled OS10k, so i would wait at least to the next maintance release (should be .679) before bringing that to production.
If you can get hold of lab release 670 you could test, i would not try with the currently available (645) build in a production environment.
With the 670 we have the two OS10k up and running, without any problems so far.
Holding about 100 linkaggs to edges
did you had more time to test the VC setup on the 10K switches as we are near to plan the upgrade in sept/oct to remove MultiChassis setup and move our VRRP setup we have from the TopOfRack switches to the more performance 10K switches and to simplify our setup of our network topology
Our problem is that we can't setup a test environment, all need to be tested onto the live systems and we would like to keep the downtime as short as possible.
We have a 10k pair as VC up and running for about 3 month at a customer site.
So far without any problems, but the 10ks are just a fat core/Distribution setup, housing some 8X vlans and 120 LACP aggs as downlinks to edges and have a simple dvmrp to have Multicast Routing between the networks.
Nothing fancy, nothing special, so so i have no experiences with ERP, gvrp, vrf aso.. your miles may vary.
Our company core consists of two 6900 as VC so far without any bigger issues..
You can just convert/merge your MCLag config to vc: create a new directory on both chassis, create vcboot, vcsetup (maybe have a example of a VC at place..) remove all MC-LAG Stuff (multi-chassis active, vip ip ...) and create lacp aggs with the proper ports (1/1/X, 2/2/Y).
Reboot from the directory and (if you did a good job) it should be fine with downtime of the reboot..
If you hav errors check whther they seem fixable in your maintanance window, if not reboot from old directory..
Use at least 7.3.1.682 (i have no experience with 7.3.2 atm).
Hi,
you can use the newest build from 7.3.2 (now it is 7.3.2.374R01). Maybe until feb there will be a new maintanance release. So you should use this one. We wait for 7.3.3 in the next days where VC with up to 6 chassis are supported. If this feature is not necessary for you use the last build 7.3.2. It works in a lot of installations stable.
regards
Silvio
And VC with 6 Chassis will afaik not be used by 10k only 6900.
If you have/plan to use 10Gig Base T cards you need 7.3.2 (at least in 6900 - no 10G-T for OS10k atm).
I think the latest 7.3.2 should be ok, there are probably some issues to be ironed out, but there is nothing that comes in my mind which is a show stopper. the next maintanance release brings commands to remove the auto-fabric generated commands which i would like, but you can do that by hand.
What i would suggest (as you do lab tests): use a 7.3.2 GA and test an issu upgrade in a OS10K-VC (we had some strange) issues in this setup but we were coming from an not so really supported release. As this was a production setup i could not make them test it again, so i am interested in your findings.
So i would suggest doing a issu when:
a) in normal operation
b) from secondary CMM in primary chassis
c) with secondary chassis active
d) from secondary CMM when secondary chassis is active.