On 13 Nov 2008 17:07:42 GMT, Jim Yanik wrote:
any speculation on what was "causing" the "bad sectors"?
Head clog. It's a direct function of the runtime of the drive (and
the number of starts and stops). Also, various chemical outgassing
and corrosion issues inside the HDA.
I've seen exactly the same thing in various RAID arrays. Near
simultaneous failures of the drives. Years ago, I had one RAID 5
array that killed 4 drives in about a week. It was pure luck that I
was able to keep the beast up while re-mirroring a drive, only to have
the next drive subsequenty fail.
However, things got worse after that. I decided to give up on RAID
and went to independent drives, with some of the drives acting as an
image backup of the main drives. They would get backed up at night.
In theory, all I had to do was remove the main drive, shove in the
image backup, and putter along merrily. I never had to do that.
Instead, both sets of drives again failed nearly simultaneously about
2 years later. Even though one set of drives was activily seeking,
while the other was just spinning, the lifetimes were about the same.
So, I thought I was safe by removing the backup drives. Different
customer, same problem. I had box of brand new drives sitting on the
shelf. One drive in a RAID 5 array was acting funny (according to
the Mylex controller), so I replaced it with a new drive, which failed
in about a week. So, I crammed in another new drive, which also
lasted about a week. Then the drives in the active RAID 5 array
started to fail. I bought some new drives from a different vendor,
and rebuilt the array, thus saving the day. With these dogs, it
didn't matter if the drives were running or just sitting in the box.
Both ways, they failed nearly simultaneously. (Maybe there's a
failure timer inside the drive).
Moral: You can't win.
Lately, I've been doing much better on the drive lifetime and RAID. I
still have some striped arrays (for speed) running. I'm using various
Seagate drives and have had no failures for about 3 years. With my
luck, they'll probably all die at once, but so far, the stats show
that they're doing fine.
Note that this is all hardware RAID. I won't use software RAID due to
some bad experiences trying to recover.
Incidentally, my office SCO Unix box has been continuously running a
Conner 1060s SCSI drive since about 1993. I keep waiting for it to
fail, so I have an excuse to build up something better. The 486DX2/66
mother board has died twice. One power supply died. Some RAM went
bad. But, the hard disk and Ineptec 1542 controller keeps plunking
along for 15 years. Hint: It's always on.
--
Jeff Liebermann
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060
http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558