Splitted live-snapshot patches + some comments on snapshot internals
On Monday 09 June 2008, 02:46:53, Otavio Salvador wrote:
> > The latter ext2 is meant to ease the creation of the "Cow" persistent
> > media instead of using tmpfs which is the original casper solution for
> > persistence from ubuntu. But this way has proven to be slow in practice,
> > although is the most "unplug power cord" friendly one.
> Could you give a better explanation how this tmpfs solution used to
> work? I didn't see it and would be interesting for a whole situation
It is easy: aufs and unionfs needs a writable filesystem to use for "cow"
(copy on write), the default debian-live boot mounts a tmpfs to provide that.
If "persistent" is a boot time parameter the script will try to find a file or
a partition named (labeled) "live-rw" and if it finds it it tries to mount it
instead of the tmpfs, so your modification will survive reboots.
This solution is expensive for a usb-key, although is fine for a mounted hd
partition or a file image. Since it was expensive for usb using, it rose the
needings for a faster system like my snapshot solution.
This solution should work right now, but live-snapshot could not help you in
building the file-image at runtime since "live/cow" is hidden 'cause of a
mount option bug.
> > Any wishlist or ideas on snapshot/persistent topic here?
> I've been using ext3 since ext2 corrupt too easily.
ext2/ext3 are both supported now.
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.