Panther and Disk Warrior 2.1

So, I have Disk Warrior 2.1. I recently installed Panther. My iTunes
is doing some weird things and I am wondering if it is time to run
Disk Warrior.

Has Panther changed around the filesystem at all/enough that I should
worry about not using DW 3.0?

Are there any other things in DW 3.0 that make it worth upgrading
(aside from being able to run DW from OS X, which has
convenience-related advantages)? The DW website is not very clear
about what 3.0 gives you over and above 2.1.

Thanks!
Bruce

 

One important thing it gives you is SMART monitoring, so it can tell if
your hard disks are about to die before they actually do.

Also, it's a whale of a lot faster than 2.1 at fixing disks.

 

While I don't know for a fact, I would assume yes, you'd need 3. With
major upgrades in 10.3, and also, depending on your computer, OS X
booting only computers wouldn't be able to boot from the 2.1 CD

I don't know if it's just my machine or not (a G4/500) but I find DW3
(and Norton 8) to take FOREVER to boot from the CD into OS X

 

2.1 should work fine with Panther disks, as long as your Mac can start
up from the CD (which contains OS 9). However, rebuilding a disk with
it will turn off journaling. Since journaling is on by default in
Panther, you'll want to turn it back on after you run DiskWarrior 2.1.

 

I wouldn't use Norton, but I did notice DW 3 takes about
a minute or 2 to start up - same as DW 2.x in OS 9.

 

No, that's not how I read it at all. HFS+ is HFS+; nothing has changed.
The only question is how you're going to start up DiskWarrior; if you
were using 2.1 before (presumably starting it up under OS 9 from a
different partition) then this hasn't changed either. m.

 

How do I turn off/on Journaling? Is that in NetInfo? And what does
journaling do?

I have been using 2.1 now and then under OS 9, which my machine boots.
It's about the only time I ever use OS 9 (or even classic, for that
matter). The SMART monitoring thing in DW 3.0 sounds useful though.
I guess I will repair this stuff now with version 2.1 and order 3.0
for future use...

I'm still going to watch this thread though, I'm learning a lot. My
sense was that HFS+ hadn't changed and that DW 2.1 was probably fine,
but I don't really know enough about what goes on under the hood of
Panther to be sure, so that's why I asked.

Thanks!
Bruce

 

In Panther, there's a button in Disk Utility that does it. Journaling
is a system feature that helps protect the integrity of your disk in
the event of a severe system crash or power failure. UNIX basically
caches everything it can for as long as it can, under the (usually
mostly valid) assumption that the system won't crash, with the result
that sometimes changes don't get written to disk immediately. Instead,
the system keeps the changes in RAM, waits for a period of inactivity,
then does the writes. If you do have a crash (by which I mean the
whole system goes down unexpectedly), this can leave your disk in a
rather confused state. Journaling writes changes to a temporary area
on your disk so as to avoid losing them. Why not just write them to
where they belong in the first place? Because that would be slow. At
least with journaling, all that stuff gets put in one area of the disk,
so your hard drive isn't seeking all over the place. So while
journaling slows disk writes a little, it doesn't slow them as much as
making the cache "write-through" would.

 

It also makes it *much* faster to recover. It is much faster to
"replay" and transactions still in the journal than it is to examine
and try to repair every file and directory on the disk (and the bitmap,
and volume header and ...).

That's true, but there is a more important reason. The HFS Plus
structures are complex, and a single change (such as creating a file)
sometimes requires coordinated writes to several different locations on
disk. If you're using a disk cache (as most systems do), some writes
may end up going to disk, while others stay "dirty" in memory. Even if
you do the writes immediately ("write through"), you could crash after
some, but not all, of the writes have completed. This would leave your
disk in an inconsistent state, quite possibly with data loss.

With the journal, all of the changes are written past the (then
current) end of the journal. After all of those writes, a single write
to the journal header changes the pointer to the end of the journal.
This has the effect of atomically adding all of the writes at once (or
not at all, if you crash before updating the journal header). After
all of the changes are safely in the journal (and the header is
updated), they can be written to their normal locations on disk
whenever it is convenient to do so.

When all of the changes eventually get written to their normal
locations on disk, the journal header is updated to change the pointer
to the start of the journal (atomically removing all of the changes
from a single transaction).

It's this all-or-nothing behavior that gets you the speed and integrity.

-Mark

 

How do I turn Journaling OFF if I want to. Once it ON the button is
dimmed out.

sam

 

Or you can customize the toolbar and add an "off" button.