Server upgrade? [update: 2012-02-17]

Ah, the life of a consummate hardware tinkerer.

Blog devotees will know that my “Home” server is the guy whose job is to run the house – in-so-much as the house can be run by a computer, that is.  More commonly noted in these journals as “Home Automation”, the Home server is primarily responsible for surveillance, HVAC, Whole Home Audio, TV show transcoding, and a host of other support services.

Said server has had a long history of service, although the only original parts are probably the case, power supply, and some hard drives(!).  And after just over 3 years with its current hardware configuration, it would seem that a significant shift is about to occur.

So significant, in fact, that I’ve had to come up with a new “P”-prefixed name for said server.  Yes, *all* of my computers are named with a “P” as the first character…

Anyhoo, this started about a week ago when I noticed some “previous system shutdown was unexpected” messages in the event log.  And some quick investigating led me to conclude that hardware was – again – the cause of the problem.

I was left with three choices:

  1. Retire the Home server and move all functionality to the remaining (“Internet”) server.
  2. Troubleshoot the problem and replace the failing component.
  3. Start from scratch with a new server

My “best practices” principles ruled out option #1.  And with time at a premium these days – and given that the Home server is based on technology that’s over a decade old – I decided to take the $$$ plunge and build a new system utilizing modern(ish) technologies.  Some things I’ve wanted for some time – like RAID1 drive redundancy – and other things I’ve had no experience with and wanted to have at least some passing familiarity (like SATA… yes, *all* of my drives are PATA!!!)

I also decided that I wanted to go with an Intel motherboard and processor solution.  In all of my professional and personal years dealing with computer hardware, I can’t recall one time that an Intel product has failed (especially motherboards).  I may be unique in this experience, but I decided to pay the extra money and go Genuine Intel.

Unfortunately I may have been too presumptuous during my decision-making process, as I opted for any old Intel board that had the hardware features I wanted (among them, 3 PCI slots) and didn’t spend too much time worrying about the “***DESKTOP***” branding that graced the packaging and product designation.  So it was that I came home, spent an untold amount of time putting the hardware together, and found myself unable to install my server software due to driver issues – all of which were arbitrary on Intel’s part, wanting their DESKTOP boards to have no business running SERVER software.

So it took a few nights and early mornings to finally get Windows Server 2003 installed and recognizing the DZ68DB board’s many features.  Thanks to Google and these pages for the invaluable assistance.

That was the first hurdle, and I knew that the software setup would be a multi-day affair.  Unfortunately I’ve run into another (temporary) roadblock, which is that the act of promoting this Windows 2003 Server to a Domain Controller has been foiled by Exchange 2000-related issues in my Windows 2000 AD schema.  Argh.  And while I value the learning experience, at this point I just want to get to the point where I can start moving services to this new “Home” server from the old “Home” server.  I’m really looking forward to the significant increase in speed that this server should bring, particularly with respect to show transcoding and database operations.

So I’ll update this post when I have more to report.

[update 2012-02-17]

So I’ve surmounted most (all?) of my AD-related issues, and the server is chugging along quite happily.  Girder gave me some issues – particularly in the area of launching processes – but this was resolved by changing the window parameter from “hide” to “show”.  Had to do it for both the open process and kill process (this is more a note to myself for future reference…)  I was also pulling my hair out over mixer issues until it dawned on me that the references to the hardware on the old server were probably causing Girder to get mired in some thick mud.

Anyhow I can say that services are smoking fast; DivX transcoding is limited by the speed of the network transfer from the PVR machine (which is currently the Internet server but will soon become the Home server).  The responsiveness of the whole-home audio interface is likewise much improved, as is the whole set of intranet pages (particularly noticeable when looking at archived HVAC and surveillance data).

I’m still in the process of moving my network-stored movie images from a local disk on the old Home server to a local disk on the new server.  This task’s painfulness is exacerbated by the insistence of the old server to spontaneously reboot whenever I try to initiate such a file transfer over the network – which certainly would seem to mimic the original problem which started this whole process in the first place.  The question still remains as to why this causes a server reboot.

Anyhow, I’m using the external 1TB drive (thanks Mark!) to sneaker-net the files from the old server.  However, we’re talking about something like 40GB of data (or is it 80GB?), which is very slow to transfer using the old server’s USB 1.1 ports…

Other than that, I’ve finally got full data redundancy thanks to Second Copy (thanks Daniel!) and that very same 1TB drive.  NTBackup is backing up the system and data drives on both servers, saving said backups to the external drive; and Second Copy is doing one-way replication to the external drive of all media on a nightly basis.  To Centered Systems – the cheque will soon be in the mail 🙂