Thanks JP. The newer code allows LAG A/A. To your last point we'll be testing Config-Sync next week.
On Sun, May 26, 2013 at 8:13 AM, JP Velders <[email protected]> wrote: > > > Date: Thu, 23 May 2013 08:44:36 -0700 > > From: manderson <[email protected]> > > Subject: [c-nsp] Nexus 5k dual sup design > > > Hello, we currently use a single sup design in one of our DC's and we're > > playing w/a dual sup design for single homed servers in our other DC. > > Cisco A/S, our SE, and myself are not particularly fond of the dual sup > > design. However, the lead engineer appears to be set on going this > route. > > > Other than having two different architectures at two different DC's, and > > designing a network to support single homed servers, I'm looking for > > additional pros/cons as points of discussion. > > Dual-homing FEX'es allows for lower impact on upgrades (and has been a > saver before, given that weird stuff happens on NXOS). > > However, a dual-homed FEX does _not_ allow for LAGs downstream (to > dual home a server in active/active mode), at least on the 1st gen > FEXes we have, and that stupidity (together with all the STP > limitations) has steered me away from N2/5K for real datacenters. > > Do make sure to look at any vPC setup with an almost pessimistic view > to work out all the failure scenarios and design accordingly. > > Also remember that in a dual-homed FEX setup you need to duplicate > everything on both N5K's, and inconsistencies can be impacting. > > Kind regards, > JP Velders > _______________________________________________ cisco-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/cisco-nsp archive at http://puck.nether.net/pipermail/cisco-nsp/
