I don't think Joe was saying that the valuable data is on your terminal server, what about an application that runs on a different server? For example, we run MAS200 which we have to perform an end of day update that does a bunch of different things. MAS200 is on a different server but if the client session suddenly went bye-bye in the middle of the update it would cause serious corruption problems. So, valuable data need not be on your TS server to be affected by something happening on the on your TS server. And do I really want 25 to 30 users to loose their sessions in the middle of creating a transaction or an important word document or spreadsheet just because I didn't put an extra few hundred bucks in the server? I don't think so So in our case any fault-tolerance I can put in the better I am and the better I sleep at night. So I am willing to pay the extra 400 to 600 bucks to put an additional HD in my servers. Most new servers come with an array controller so the cost is very minimal. Farms are great, so is imaging, but they were not made to protect against hardware faults. It comes down to each situation, in my case I have reasons and prejudices that make me want to maintain RAID on my servers, others may not. Jeff -----Original Message----- From: Jim Hathaway [mailto:JimH@xxxxxxxxxxxxxxx] Sent: Friday, July 12, 2002 12:32 PM To: 'thin@xxxxxxxxxxxxx' Subject: [THIN] Re: What is the optimal RAID configuration valuable data on a terminal server? surely that should never be the case. blood, sweat, tears poured into creating your config, yes. Besides, with imaging, your downed server is only down until your new drive arrives. A load balanced farm does mitigate the risk of loosing any one to two servers out of a farm due to hardware failure quite a bit. To be fair, performance at Raid 0 on most drive systems these days is only slightly better than Raid 1. Hardly noticeable to end users and perfmon.