I'm curious to know if anyone is using and implemented anything with RSMLT yet, and how it's working for you?
Currently, I have about 20 8600's world-wide running, and 4 of them are in a full mesh in my network core, doing SMLT with VRRP. All are running 3.7.4. I've been toying with the idea of taking the 4 in my network core and converting the VRRP SMLT's for linking the full-mesh together, and simply making two groups of two 8600's, with an IST between each group of two, and then just DMLT's between the two-node groups, using RSMLT (creating a pure square design instead of the HA full mesh.) Couple that with HA-CPU mode and it would be pretty cool for uptime.
I haven't had time to play with this in my lab yet, but any real-world gotchas are always nice to hear!
Thanks!
Currently, I have about 20 8600's world-wide running, and 4 of them are in a full mesh in my network core, doing SMLT with VRRP. All are running 3.7.4. I've been toying with the idea of taking the 4 in my network core and converting the VRRP SMLT's for linking the full-mesh together, and simply making two groups of two 8600's, with an IST between each group of two, and then just DMLT's between the two-node groups, using RSMLT (creating a pure square design instead of the HA full mesh.) Couple that with HA-CPU mode and it would be pretty cool for uptime.
I haven't had time to play with this in my lab yet, but any real-world gotchas are always nice to hear!
Thanks!