Livecd from ram

[Martin Meredith]

Usually though - isnt it usual that the amount of memory/amount of swap
space available is about right to be able to run the programs... ok
accessing stuff from hard disk and ram might be quicker - but you have o
take into account the fact that by using swap space or ram - you're
detracting away from that ram/ whatever - so things are likely to go
more slowly - ok - the inital load may be quicker depending on
configuration - but i'd rather have a slow load time than an
unresponsice application

Phillip Susi wrote:
> I've been kicking this idea around for a while now in my head and now
> I'd like to get some more opinions.  I think that the livecd would
> perform much better if the cloop/cramfs image is copied to a tmpfs at
> boot time and mounted from there.
> 
> What about systems without enough ram to hold the whole image?  The most
> common use cases of the livecd will be in a system that has a hard drive
> that already either has a linux system or windows installed.  If the
> system already has a linux system installed, there will be a linux swap
> partition that can be used.  If windows is installed, then I thought
> hey, windows has a pagefile already on the disk, simply using that
> shouldn't be too hard right?
> 
> The initramfs on the livecd could first probe for a linux swap
> partition, and if not found, then check all fat and ntfs partitions for
> a windows pagefile.  Even if it is on an ntfs partition, you can still
> get the block list that the existing file lives in and bypass the
> filesystem to swap to the pagefile.  You would need to skip the windows
> pagefile header, but the rest of the file could be turned into linux
> swap space.  This could be done with a small utility to get the block
> list and set up a kernel device mapper device mapping the pagefile, then
> mkswap and swapon that dm device.
> 
> With some usable swap space, even mid to low memory systems could copy
> the livecd image to the swap space and then running from there should be
> much faster than from the cd.  Of course, in the event that no swap is
> found, the system will run like it does now, and the user can use a boot
> parameter to force one way or the other.  What do you guys think?
> 
> 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 256 bytes
Desc: OpenPGP digital signature
Url : http://lists.ubuntu.com/archives/ubuntu-devel/attachments/20051225/31974089/signature.pgp