|[Ben]:||A tale of two servers ||Discuss This [0 comments so far] View Comments|
|So what happened?|
The problems were many and varied. It started with repeated power losses that caused the old server (which never had a good time starting up) to stop locating the B channel of the SCSI adapter. Unfortunately for me, that's the channel the RAID array was parked on.
Okay, so what did you do first?
I run Windows SBS backups on the server every other day or so, so I figured I'd just fire up Windows on a brand new server (and quite an upgrade at that) and restore it. Unfortunately, I learned that Windows SBS backup requires restoration to identical hardware. What the heck good is that? Oh, sure, if I accidentally delete a file, I can recover it, but so frickin' what? When I do a backup, I want to do a backup. Something that will let me kick-start a new machine (I don't care if it is a calculator or a Cray), pour in the backup and be back up and running. In short, I want an Acronis backup.
Wow ... so you couldn't even restore a backup?
Nope. Well, not on new equipment, anyway. And my server was an ancient IBM Netfinity 5000 - where do you find one of those on short notice? The answer to that is "In the back storage room of a local business." My dad had purchased four of them for cheap years ago. I was able to cobble together a working box from those four old servers, slid the old drives into the new machine and fired it up.
Okay, so then what?
When it came up, I was in familiar territory. The RAID card was reporting that every drive in the array (except the hot spare) was defunct. DDDs all around.
I repaired that error in no time (having done it often enough in the past) and sure enough, Windows fired back up.
So if you fixed it so quickly, why did it take so long to come back up?
Hold on, I'm getting to that. When I brought Windows back up, I was getting a boatload of error messages complaining about services not starting. It seemed that every single Windows service necessary to run the website from Active Directory to IIS to Exchange was failing upon startup.
Well, at least I could now restore from a backup. So I did.
And then it worked?
No. It didn't. The restore process turned out to be utterly worthless in this case.
So, I completed a repair installation and things seemed to work better after that. Then I made an Acronis image of the server and put her back online.
So you can restore the site if it dies again, right?
Better - I'm actively looking for new hardware to put the server on. I'm probably going to go with an HP ProLiant DL380.
You didn't answer my question. If the site goes TU again, you can fix it quicker, right?
So why did you have to go through this whole thing? That's all I wanted to know.
Then you asked the wrong questions. And you say "so" too much.
A tale of two servers