Doing workstation maintenance for a client today, I found there had been a power outage earlier. Checked the servers, just in case. Prevents me from being paged in the morning if anything went awry.

Two of the servers rebooted because they lacked UPS power to go through the entire length of the outage. Made sure they came back up okay, but it turned out the Exchange server didn’t. This is the second time I have ever heard the horrible sound of an alarm indicating a dying SCSI drive as part of a RAID 5 array.

Last time was on a machine under warranty. This time it’s a machine just under six years old, nowhere near under warranty, with SCSI drives in a size I suspect will be hard to get if I merely want to swap out the bad one and maintain the status quo until the hoped for “upgrade everything” project next year.

I disabled the alarm, and sent out a couple of slightly panicked e-mails about the problem. Which is more of a potential problem, in that it affects nobody unless another of the three drives fails before something can be done.

My current thinking is to buy a server, install the old OS and Exchange, transfer the data, then upgrade what’s on the new server next year as part of the “upgrade everything” project next year. I doubt I will be able to find a matching drive, or that it would be worth doing so.

The scary part is I will be away Thanksgiving weekend, and there is no way a replacement server will be here before then anyway. Yikes. Of course, if the client would stand for it, I’d have had them upgrade at about four years old, rather than wait for failure. Oh well.