It has been a project of almost 4 months, but the end is near!
Probably it can be done in a much shorter time span, but we took our time for this one.
Mail is still considered as one of the most vial elements in a modern company, so we were not allowed to make stupid mistakes!
If you never heard of Exchange, you’re probably not an IT-er, or an IT-er living under a rock. All communication, reminders, calenders and contact information is stored in this massive software package from Microsoft. (Sometime users even use this to archive their pictures, music or whatever data they have… 😮 )
And you don’t want to delete random things, or run into corruption…
So we really double-checked (and even tripple-checked) every step we took!
The core of the story: moving >1000 users from exchange 2007 to a complete new, high performance and redundant exchange 2010 environment.
So a couple of new servers where bought: some nice dell poweredge’s featuring a AMD logical octacore processor, 64GB of DDR3 ram, and a bunch of sata drives.
Exchange 2010 is less depended on raw iops and cpu cycles then exchange 2007, so we made a nice compromise between price and performance.
On this hardware, another scoop was introduced: the first production hyper-v solution in our server park! (all of our current clusters run vmware esx 4.1 and 5)
So we have 2 new poweredges with hyper-v core, both with 2 guests for hosting the different exchange server roles, and a couple virtual machines for authentication and reverse proxy!
Redundancy is created easily using the new Database Availability Group (DAG) feature of exchange 2010.
This runs on top of the windows server clustering technology, and it mirrors the information across servers.
As a result, information like your mailbox/calendar/contacts is, pretty much, always available to the end user. When one of both servers explodes and goes down (which is not unthinkable) or a fiber is cut, an airplane crashes onto our building, …, the other server takes over.
More redundancy is created on the client access servers and the internet-facing reverse proxy servers (using Microsoft’s Forefront TMG), which all are in twofold.
So try taking that down! 😛
(Actually, mail went black already twice… :-?, but not Exchange or hyper-v is to blame! )
Most of my time went in learning Exchange Server, preparing for possible pitfalls and installing TMG (which was a real challenge, maybe I’ll post a real tutorial for that later).
I made and reused some scripts in Powershell for managing transitions, monitoring and other purposes.
I’ll definitely post them later, because I’m sure you can learn from them the way I did!
Exchange, over and out!