After the initial design thoughts on our HyperV solution I got some quotes in for equipment and worked out how many servers we could probably afford this time round. Happily it is more than the original 3 I had set out my design on, so here is an updated design based on 5 HyperV server and 1 SAN
Then it was a case of working out how much memory I could use out of the 72Gb I had specified on the servers. I thought for good measure I would reserve 8Gb for the actual HyperV server, leaving 64Gb to be shared out among the virtual servers. Now I am not very good at maths so please feel free to ridicule my sums.
- 72Gb – 8Gb = 64Gb for virtual servers
- 64Gb divided by 5 = 12.8Gb (leave this amount free for failover of any ‘dead’ HyperV server)
- 64Gb – 12.8Gb = 51.2Gb (left to share among virtual servers on single box)
Now this is only a very basic working out, more detailed math needs to be done to make sure I get it perfect.
This image shows the current and envisaged new servers that will be hosted on the HyperV environment. Even if I assigned 8Gb to each virtual server this would still leave me with 10Gb free on each HyperV box and therefore the ability to install an extra 5 servers over the whole HyperV environment.
So that’s where we are at so far, the next phase is the actual ordering of the equipment and then finding somewhere to put it to build it up – you have to remember that I already have a rack full of physical servers so no free rack space to put 5 new servers and a SAN, but its a problem I can live with.