Hello everybody,
We've been running ESXi 4.x for years here, it's been rock solid.
Well, the time has come to move to 5.1 and needless to say, it has gotten WAAAY more convoluted! After much pain, I have a new 5.1 vCenter up and running (and all of it's components). I still have our old vCenter 4.1 server up with some older Sun/Oracle machines and two Dell R910s as the hosts. My new VMware environment consists of 2 new Dell R910 servers to replace the old Sun hosts, plus I want to move the two R910's that are currently on my 4.1 vCenter to the 5.1 server and upgrade them to ESXi 5.1. (They would be two separate clusters.)
My first stumbling block (after getting vCenter 5.1 up and running) is cleanly adding the brand new R910's to vCenter and have the new vDS working properly. Unfortunately, with all of the documentation out there, I cannot seem to find a straightforward "How-To" for the most basic of setups. :-(
Our R910's have 6 NICs (4 Broadcom 1Gbps, 2 Intel 10Gbps). We currently have a stack of HP 2810 1G switches, but will eventually move to Cisco 3750 switches.
I set up 4 portgroups in my vDS (one each: PG-Vms, PG-FT, PG-vMotion, and PG-Mgmt) for each of the different "types" of traffic. Do I need to associate the different "profiles" to the different portgroups? (If so, how do I do that?) Should I restrict the individual ports (upLinks) to the different port groups, or do I just let the intelligence of the vDS do it's magic? Can I team a 1G and a 10G nic together and get 11G (with failover) or should I just set it up with the 10G as primary with the 1G as secondary? Is it a bad thing to plug all of the NICs into the same switch or stack? (I have heard conflicting stories...so it really isn't as dumb a question as it sounds.)
I am going to the official vSphere training, but that isn't until March...and unfortunately, I need to get these clusters up and running now. :-(
Any help y'all can offer would be GREATLY appreciated! (Even if it's as simple as "here is the link to the How-To doc, you big dummy!") :-)
Thanks,
Emmett