We're a dental practice currently running Windows Server 2003 on a Dell Poweredge 1800 which has 4 160GB SATA drives in a RAID 5 configuration.
Can partitions be safely resized on a RAID 5 setup?
The reason I'm asking is our dental practice management software has a new version out, but I'm reluctant to deploy the upgrade until I can be sure that said upgraded software will run acceptably under a Windows Terminal Server environment. I've done a trial upgrade on a copy of our database (on another system running a trial version of Server 2003), and accessed that trial upgrade remotely via Terminal Server. For the most part, the upgrade completed successfully, and the software runs.
However, screen updates in the appointment scheduler portion of the software are painfully slow over a remote connection. I'm trying to find out whether it's just because the trial system is underpowered (it is in at least one respect: the software requires 2GB of ram for the server, and the trial system only has 1.5GB installed), or whether it's because the upload portion of our DSL connection is just too slow for the way the new software draws the screen in the scheduler.
To that end, I would like to find some way to a trial upgrade on our production server, but in a dual-boot setup. I'd adjust the size of the partition on the server so I'd have room for a trial copy of Server 2003. I'd install the trial copy of Server 2003, and then install the new version of our software. When I was ready to test the new software remotely, I'd set up the boot partition to boot the trial version, and do my testing (naturally, during a time when we're not using the production database.) When testing is complete, I'd change things back to the original boot setup.
I just loathe the idea of going ahead with the upgrade on the production server, and "hoping for the best", because once we upgrade the database to the new version, it's AWFULLY difficult to revert to the old version.
Please point out the flaws in my line of thinking. I'd like someone else's thoughts on whether the potential reward (testing the new software and finding out whether indeed it *is* or *is not* usable via Terminal Server) is worth the potential risk to the boot record on our production file server.
I may end up just not doing the scenario I'm describing here, simply for the fact I'm not wild about the idea of monkeying with the boot setup on our production file server.
Of course, I'd have at least two backups of our database before proceeding with *any* testing.