| Re: .0
The disk validation process is part of Amiga DOS and runs even if you
never do a LoadWB.
However, when the workbench starts, it opens every disk to see if there
is a custom disk icon (file disk.info) for that disk. That could account
for the random disk reads when workbench starts. If you open lots of
files or do lots of logical name assigns, the workbench will even prompt
you to insert disks in drives so it can try and read the icon information
for the disk.
Thirty seconds seems a bit long for this process to continue.
|
| The procedure I used for startup is the same as
what I used for other releases of 1.2, it's only Gamma1
that does this. That's why I wondered if anybody knew
(heard) if gamma1 was peculiar in this respect.
(If so, I have motivation to get the official 1.2 release.)
Dave
|
| I once got a lousy copy of a 1.2 field test release, and experienced
pretty much the same problem. While I never found what theproblem
was precisely, I figured out the following scenario which at least
seems consistent with what's happening.
When you put a disk in a drive, it runs the disk validator to make
sure it's sane. Actually, I discovered that there's really two
pieces of the validator, which for the purposes of this discussion,
we'll call "little validator" and "big validator".
When you put a disk in a drive, it runs little validator to do the
sanity check. If everything looks ok (which it usually does), it
finishes in a couple of seconds and goes away.
If everything isn't ok, it'll run the big validator. This isn't
usually resident, and will have to be loaded from your "system"
disk. (Some people have noticed that on a single drive system,
the mere insertion of a hosed disk into an idle system will nearly
immediately cause a requester for the system volume. This is why).
Big validator will run anywhere from a few seconds to almost a minute
depending on how much stuff there is on the volume. Its main aim
in life is to locate the error and to repair it. Once it's done
so, it goes away FDH (fat dumb and happy).
The problem that I was experiencing with this bad beta release was
that there was a file system corruption on the workbench disk, which
got faithfully reproduced when I DISKCOPYed it to my working copy
(since it was a "soft" error rather than a broken sector). Thus,
when I boot the disk, I get big validator shortly after the system
boots.
The only question is, why did it happen every time I booted the
system? Well, because I had the disk write locked. Every time
I'd boot, it'd run big validator, who would eventually find the
problem, then fail to fix it because the disk was write locked.
The solution was simply to insert the disk once while write enabled.
After that, no more problems, and it took considerably less time
to boot the system.
|