[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: MFM drives / IDE drives RTBF



Dear Jan,

The usual "failure mode" for MFM drives is that the head positioning
mechanism in the drive wears and changes the relationship of the heads to the
tracks on the disk.  This makes it difficult to read what you wrote.  If you
reformat the drive, you correct the track-to-head relationship, and it works
again for a while--though maybe not for long if the mechanism is getting
sloppy.  Unlike a PC running a standard OS, you can't just reformat a daVinci
drive because a lot of the underlying software you'd need (Regulus and so
forth) is factory-only stuff that end users don't get.

Modern IDE drives use a voice-coil head positioner that is much less likely
to get into trouble with use, so if an opportunity comes up to replace an
old-style drive with a new one, it's worth doing.  Unfortunately, a new drive
controller is needed, and that's not  simple to change in a daVinci.

I do not understand why your drive supplier says that MFM drives don't last
as long on a UNIX/Regulus system as on a DOS system.  The two are actually
pretty similar in how they treat hard disks.  What he may be alluding to is
that in an industrial/commercial environment, where Regulus is normally used,
machines tend to be in use constantly--as opposed to a PC sitting on
somebody's desk at home that only gets turned on for a few hours every
evening.

We have three da Vinci systems here (a "classic" which was retired just over
a year ago, the 8:8:8 which replaced it, and an analog Renaissance), and have
experienced no hard drive failures in the past five years.  These systems are
in use 20 or 24 hours a day, five days a week.  The important thing to know
about da Vincis and hard drives is this: whenever somebody is logged into the
system, a daemon loads which causes the system to save all session settings
every ten minutes.  This happens whether or not anybody is actually doing
anything.  If the machine is left logged in for a week, more than a thousand
disk writes--most of them needless--will occur.  Over the course of a year or
so, you can see how this adds up to a lot of wear and tear on the drive.

It is very, very important to make sure all users log off the system when
they aren't doing a session!

Another reason for logging off when you're not working is this:  if a power
glitch happens during one of those automatic disk writes, there's an
excellent chance it will scramble something on the disk. 

If your users are logging out properly and you still seem to be getting short
drive life, you might want to consider the possibility that you've got a bad
drive controller.  Wear in MFM drives is compensated for--up to a point--by
the error correction inherent in Miller-FM coding.  The error handling loop
in the drive controller circuitry could be defective, which would allow a
newly formatted drive to work for a while.  Another possibility is that
something in the controller is slamming or oscillating the head positioner,
causing it to get out of alignment prematurely.

Good luck,

Christopher Bacon
DuArt Film & Video