Issue 5

Let me tell you a horror story. A number of weeks ago, I set out to Raid two 4.3giggers on my Mac, using Conley's softraid on an Initio Miles controller. So, as a wise, experienced network admin I set out to back up everything before converting. So, we throw about 6 gigs of data on the NT Server, and began backing up to tape. "I don't need to wait for it to be finished," I bravely declared, and began the Stripe in place process of Softraid (Hold the comments for now ;)

About a 1.5 gigs into the backup process, I started to get errors writing to tape, and errors reading the volume. This isn't good, I thought. So, I figured, I have some free space on another machine, so I started copying the files onto it from the server.

By this time, the Mac was crashing every 30 seconds. For whatever reason, my system DID NOT like Softraid. I was beginning to panic, but figured that the volume will be fine once the stripe in place process completed, and I just needed to keep restarting the Mac every so often. Looking over to the other client, I noticed that I was getting read errors from server.

Uh oh.

The backup process pretty much stopped at that point. I grabbed what I could copy from the NT box, and rebooted the server.

The server was gone.

Assessing the situation, I decided it would take me about a day to rebuild the primary partition, and I could attempt recovery. I didn't have anything on the server that wasn't backed up (except the 6 gigs fresh from the Mac) and besides, the data is still on the Mac.

Or so I thought.

NT refused to reinstall. No matter what I tried. At that point, I pretty much understood that some piece of the hardware has gone bad, and at that stage, I just figured it was time to build a new server anyway. The Mac completed the stripe in place process- and the volume became completely inaccessible.

Nothing that I've ever experienced has prepared me for this kind of disaster.

It took me another 2 weeks to order all the parts to build me new server (which rocks, BTW), Installed NT- And the data partition, which I didn't touch on the drive is gone. Like it never existed.

Long story short, we had most active files either on current zip disks, and some I managed to copy before the server died. Much of the older files copied were previously backed up, so all in all the data loss was not nearly as great as it could have been. We lost about 8 hours of work we needed to reconstruct, and some otherwise unimportant program and data files. Today, I have a brand new, dual P2 server with 16 Gigs of space, dual UWSCSI, complete fault tolerance, and scheduled backup on a 32gig autoloader. The Mac remains RAIDless, and I suspect that will not change anytime soon. The folks at Initio expressed their regret for my experience, and offered a new BIOS.

As of right now, I'm still trying to figure out what the moral of the story is. I think its something like "All computers are evil." When I recounted this story to my friend, who is a network admin in Silly Valley and a devout cultist of Linux, he simply offered that none of this would have happened if I used a real Operating System. "What Operating System would you suggest?" I asked, fully knowing to expect the onslaught of technobabble coming.

Next Issue- Linux: the preferred server of choice in the multiplatform environment?

I welcome all questions and comments at akoyshman@jps.net or designamics@jps.net