Lauri Kasanen wrote: > can someone tell me how is the device mapper used in the lfs livecd?
device-mapper is used for pretending that the CD is fully read-write. This is done with the help of two loop devices. /dev/loop0 is set on top of the root.ext2 file, /dev/loop1 is set on top of a huge (1.5 GB) but sparse (i.e., zeroed-out regions don't take space) file in /dev/shm. Then a "snapshot" device-mapper target is used to combine the two loops into a /dev/mapper/lfs-cd device that has initially the same content as root.ext2, but can be written to (changed sectors go to /dev/loop1, i.e. to tmpfs) > I'm also interested in why was it chosen. Another alternative is to use unionfs in order to pretend that the CD is fully RW. The 6.1.1 CD did this, but it was easy to cause a kernel oops or a full system hang. unionfs has been removed exactly because it was too buggy and unstable. > At least on my older > laptops it tends to cause some "block reading errors" that fill the > screen, and then the system hangs. On the other hand, cloop-based > livecd's work fine on these lappies too.. Cloop was never used on LFS LiveCDs. Older CDs used squashfs for compression, while modern CDs use zisofs. The change was done because, with device-mapper based RW mechanism, one has to compress a single file (root.ext2), not a full filesystem tree, and both zisofs and squashfs are equally efficient for that - but zisofs is simpler to set up. BTW, this question is more suitable for the [EMAIL PROTECTED] list. And please post the errors from the latest LFS LiveCD (r2052 or self-built from SVN trunk) - they are bugs. -- Alexander E. Patrakov -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
