Re: Fatal: Out of memory

[Evin Robertson]

Andrew Clausen <address@hidden> wrote:
[...]
> Also, Parted can't grow the cluster size.  So, if you have
> a small image with a small cluster size, then that means a big
> partition will have a LOT of clusters, because the cluster size
> is small.

Hrm, I'm thinking that even if I could get enough memory/swap
available for parted, then windows itself might have problems dealing
with so many clusters.

>> I obviously don't have GDB on my boot floppy, or I'd try to debug this
>> myself.  (My experience is that something gets upset if I try to munge
>> partitions on a drive on which I have a mounted partition, so
>> installing linux and running GDB that way doesn't seem easy).
>
> It shouldn't matter... as long as:
> (a) you don't touch partitions that you have modified
> (b) you don't modify partitions that you are touching

When I had a partition mounted on the same disk as I was doing the
resizing, parted warned that the kernel was unable to reread the
partition table, and further parteding did bad things.

>> Alternatively, is there another way to do the equivalent of "sys c:"
>> from linux?  That's the only real point of using the small fat32
>> partition I made.
>
> I can't think of any.

Are the locations of the fat32 boot sectors easy to calculate?  Could
I dd them, copy io.sys, and expect things to work?

Is it as simple as copying from and to /dev/hda1
bytes hidden*sector_size to (hidden+3)*sector_size ?

------------------------------------------------------------
--== Sent via Deja.com ==--
http://www.deja.com/