Ron, You rock, dude! I have finally received all the "pieces" of the puzzle and have submitted both a design specification and a shopping list today for approval. It looks like the approval is "almost" a slam dunk (there was a little reluctance to purchase a new cluster when I had previously indicated that I "thought" we might be able to take an existing cluster and dedicate it to this project... me and my BIG MOUTH again! <g>). The costs will be about $180,000 BEFORE the discounted pricing but that includes an additional BL20p for testing (that may end up falling out before it's all said and done). I can hardly wait to get the new hardware in and start setting up a new MF XP farm! MF XP has been out now for at least a few years and I have yet to get my hands in it! Soon and very soon.... Thanks to all that offered advice and recommendations (Ron, Magnus, et al). I will also post the performance testing and usability we achieve on the HP BL20p blade servers as there seemed to be some concern in an earlier thread over the sensibility of implementing blade servers. We will be deploying MF XP servers and hopefully a cluster on BL20p blades. I can tell you upfront that we can achieve sixteen 3.06GHZ processors in 6U of rack space requiring less power and air than the same number of CPUs using the DL line. Time will tell if this will be the right choice or not. We seem to be happy with our 16 BL10e blade servers that have been deployed into production for some time now. With the "Rip and replace" functionality available using the RDP server, it seems the perfect configuration for MetaFrame. We'll see... Thanks! Rob Slayden 24 Hour Fitness, Inc. -----Original Message----- From: Ron Oglesby [mailto:roglesby@xxxxxxxxxxxx] Sent: Tuesday, September 09, 2003 5:52 PM To: thin@xxxxxxxxxxxxx Subject: [THIN] Re: Data Collector recommendations Comments in line Ron Oglesby Senior Technical Architect RapidApp Office 312.372.7188 Mobile 815.325.7618 email roglesby@xxxxxxxxxxxx <mailto:roglesby@xxxxxxxxxxxx> -----Original Message----- From: Rob Slayden [mailto:rslayden@xxxxxxxxxxxxx] Sent: Tuesday, September 09, 2003 5:49 PM To: 'thin@xxxxxxxxxxxxx' Subject: [THIN] Data Collector recommendations I have yet another design question as I am trying desperately to come up to speed on MF XP and its unique design elements. I believe I have all the elements for this small MF XP server farm determined with one exception: the data collector. I am looking through the Citrix doco and even Doug Brown's EXCELLENT MIAB 2.1 doco and I see no real hard-and-fast rules for the data collector. I see the recommendation to dedicate one DC for every 50 MF servers but at what point should this server become dedicated? RON: Way before 50, for sure is you are suing published apps and not desktops. My general rule of thumb is in the 10-15 range if using all published apps with a high app turnover. Not needed badly at that point but at 20 I would for sure have one. We dedicate a master browser to this function in our 27-server MF 1.8 farm. This is a Compaq 1850 doing nothing but functioning as the Master Browser. We also have 2 full-on production MF servers assigned to the Backup Browser funtion. Given that the new farm will be 5 servers running Windows 2003 and MF XP-FR3 with 300 users and 2 published applications, can I get away with using one of the production MF servers as the DC? Yes, 2 published apps and 5 app hosts. No problem. Also, what about a backup DC? This is set by the preference level in the zone tab in the properties of the farm node. One Most preferred, one preferred, and the rest default in a zone. Does it work the same as the old MF 1.8 where you simply click a radio button and now you have a backup DC? See answer above Thanks again for the advice! Rob Slayden 24 Hour Fitness, Inc.