Comment 5 for bug 474258

I sort-of agree with everyone here...
a) I have a 100% raid server running karmic (without cryptswap). I don't trust automated tools on this system and always verify that /etc/fstab points to the correct swap md partition after any significant change and before I reboot.
b) I have a laptop with hardy, jaunty and karmic installed. The swap partitions for hardy and jaunty are defined in fstab by uuid, so they cannot destroy the wrong partiton. The karmic fstab points swap to /dev/mapper/cryptswap1 and /etc/crypttab points to /dev/sda10.... DANGER WILL ROBINSON!

I don't think this implementation is satisfactory for "ordinary" users. Why doesn't an existing encrypted swap partition have an externally visible uuid? That would solve the "ordinary" user issue (I shudder to think of the pain this "bug" is going to cause if it left as-is) in an analogous fashion to the non-encrypted swap case. After all, some of my partitions have a uuid wrapper for the md superblock, then a uuid wrapper for an lvm physical volume, then one for an lvm lovical volume which holds an ext3 file system that holds the user data.

Is it possible - I don't know enough to know how to do it.