Re: cluster size

[Andrew Clausen]

Ian wrote:
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Sorry it took me so long to get back to you.
> 
> I pulled out a 30GB hdd that I had sitting around... well, I did the
> mkpartfs on it, and sure enough, booted to windows (the new drive was
> secondary in this system).  Well, first I went directly to DOS and did a
> DIR, kinda surprised to see that the 30GB only had 20gb free.  Ran
> scandisk for DOS and it didn't work too good.  It found lots of lost
> clusters and claimed that there were too many files in the root directory
> to write them as files.

Ouch!  I suspect it's doing the same silliness as with fat12/fat16, of
ignoring what the boot sector says.  This is really bad news... I didn't
know it did this.

So, I would REALLY appreciate the error messages it's giving you.  Also:
when you run chkdsk, can you write down the number of clusters Windows
*thinks* you have, and the cluster size it thinks it has...

>  Trying scandisk in windows just complained about
> not having enough memory.  I'm guessing the MS is assuming the number of
> clusters will not go past a magic number, and running 4k cluster on a 30GB
> drive easily passes that number.  Wish I would have grabbed the scandisk
> log, oh well, perhaps I'll do that today.

I suspect you are correct.  This is the way it works with fat12 and fat16.
 
> As far as what sizes get what cluster size, I guess that would have to be
> trial and error.  Perhaps there is a system to it, like 4k=8gb 8k=16gb
> 16k=32gb?
> 
> I'll do more research if I have time today.  Let me know if there is
> something else I could try.

Thanks :-)

Andrew Clausen