Pressler2904 Jul 25, 2008   5:03 PM GMT

Tom - Thanks for the link and information - it's exactly what I needed. FYI, this is an "inherited" setup. If I was performing the RAID config for a DC (and ALL this server does is act as a DC), I would never have put it on a RID 5 array: I would have used RAID 1 or 10. Furthermore, if I was physically setting up the system, the server would have been a 1U box just like the others, and not a tower server which got shoved into place diagonally, requiring the removal of either our segment firewall or our (live) Oracle database server to access the chassis for a simple HDD replacement and RAID rebuild. And (as you rightfully point out), the addition of a hot spare is a real no brainer, especially since there are three remaining open ports on the controller.....

2,190 pointsBadges:

Pressler2904 Jul 25, 2008   5:06 PM GMT

Pardon the multiple postings... The same rack as described above holds our Citrix farm (two servers, with a third failover server here at our main location). Although I'm pretty sure I know the answer th this next question, I'll pose it anyway: Given the scenario described above, what are your opinions regarding promoting one of the Citrix servers to a DC? As I said, I'm pretty sure of the answer, but thought I'd ask anyway.....

2,190 pointsBadges:

Michigan Jul 28, 2008   12:16 AM GMT

No do not use a Citrix server for a Domain Controller. Have you ever thought of using virtual machines? Running more than one server on a VM just may be an answer for you.

185 pointsBadges:

Pet geek Jul 28, 2008   2:33 PM GMT

I'll second both answers - I wouldn't suggest promoting either system to be a BDC. For a short-term fix, I'd find the most robust spare PC you have lying around, add some memory, and then install a hypervisor. Build a virtual machine as a BDC. It won't run as well as it might, but it will work while you address the main issue. Once I had that running, I'd wait until whenever the slow time is in your organization, shut down the failing machine and fully remove it from the rack. (If possible. It sounds from your response below that the only issue is that you can't work on it in the rack, yanking it sounds like it should still be possible.) Personally, I'd replace it with the right solution - a 1U rackmount system as you mention - but in an inperfect world I'd stuff the current box full of hot spares after my request for a 1U machine went down in flames. If you think you'll get only a little pushback on a 1U, build a business case for what you could do with the tower - maybe it could run a hypervisor and be a quick-cut stand-in for a failed server as part of your Disaster recovery/Business Continuity plan... -PG

100 pointsBadges:

Pressler2904 Jul 28, 2008   5:22 PM GMT

Thanks for all of your input. You all concur with my initial opinion that the two roles should be separated... When the opportunity presents itself, we will be making a trip down to Raleigh, NC (from mid-state PA) to perform the maintenance work and kick off the rebuild. It's a long, inherited story, and the hosted rack is not configured the way either my boss or I would have configured it, but you work with what you have (of course!). Thanks again, poppaman

2,190 pointsBadges: