- Making snapshots of an entire machine. That's not as troublesome because
the entire context a machine is involved in is somewhat volatile anyway.
Think of networking - packets, even connections may be lost at any time
and reasonable software is capable of recovering from that kind of
inter-machine problems.
- Create a library that contains all the code required for a process to
make a snapshot of itself / it's pgrp etc. This eleminates a large
number of problems:
- Snapshots will only be taken at apropriate points of time.
- Snapshots will contain all the required context. Think for example
of a file of n bytes size, the filepointer possitioned to n. While
the snapshot is sitting on the disk the file now grows by some more
bytes. When restoring the snapshot an external program will not be
able to automatically know if the right position to seek to is
position n or the last byte. Think of databases and logfiles.
Ralf
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu