At 15:51 Uhr +0200 19.04.2004, Michael Schmitz wrote:
Nope, pmdisk doesn't need exclusive access to the first swap partition, just enough space to save kernel state (during suspend, all user processes are frozen and swapped out, so memory used by processes doesn't count here).
Ah, this means, user processes are swapped out to the swap "as usual" which means to the swap with the highest priority first, and the swap (the first one) then contains both "normal swap" and "pmdisk-saved kernel state" intermixed? I see, that makes a difference :) (I somehow assumed the processes were all collected together and a completely "new" swap would be written.)
> (Well maybe LVM would work?, this would help me merge the two swappartitions in that case.)I'm not sure you can use LVM 'devices' for swap.
Well, LVM logical volumes *can* be used for regular swap.
That might make a difference. I had no luck with the latest mol packages on 2.6 kernels yet...
I could send you the code (source+binaries) I am using if you want. Note that I build on woody.
But see also the mol mailing list. BTW I've looked over my mol archive: I did post at feb 1 2004 asking about a compilation problem; it's the thread with the subject "MOL release candidate available". (I could tar the archive up for you) well, it's available from http://lists.maconlinux.org/pipermail/mol-general/ and my posting is at http://lists.maconlinux.org/pipermail/mol-general/2004-February/002644.html
Christian.