crypto-loop fs made on 32 readable on 64 bit?

Tracy R Reed treed at ultraviolet.org
Wed Aug 22 05:30:40 UTC 2007


Through a series of unfortunate events I have lost my important
encrypted volume. The latest version of it was destroyed. A somewhat
older just deleted copy was undeleted from a server using debugfs on
ext2 filesystem. I also have a quite old copy archived on DVD using the
mondo backup program.

The undeleted encrypted filesystem was made on a 2.4 kernel box under
RedHat 7. A very old system. It was encrypted using the blowfish cipher.
I would provide exact kernel version etc. but I don't have the box
available to me at this instant. I can provide those details later if
necessary.

When I try to remount the file with -oloopback,encryption=blowfish I get
this error:

mount: wrong fs type, bad option, bad superblock on /dev/loop1,

Question: Will debugfs refuse to undelete the file if any portion of it
has been overwritten by other disk activity? Or will it recover a
partially corrupted file and not say anything? The system was run in r/w
mode for a short time after the file was deleted but it was idle and
there was loads of free disk space. I am hoping the odds of part the
file being overwritten are slim. Unfortunately the encrypted version
looks like entropy so unless I go through all 2G of the file and look
for something obviously non-random I cannot know if the file has been
corrupted or if I am just doing something wrong.

I copied this file to another, much more modern system, FC3 running
kernel 2.6.10-1.741_FC3smp and tried to mount it after loading the
blowfish and cryptoloop modules and got the same results as above. Not
good. So I decide that maybe my bad luck is holding and the file really
did get corrupted somehow.

So I pull out an old set of backup DVD's made using Mondo and restore
the file from the DVD's. I try mounting it and get the same error as
above. I am positive the image on the backup set is good. So hopefully I
am just doing something wrong during the process of trying to mount it
but I don't know what it could be. I have always used blowfish
encryption for this and I have used the same password for the volume
(for better or worse) for years and it is programmed into my fingers.

Any ideas?

Thanks!


-- 
Tracy R Reed                  Read my blog at http://ultraviolet.org
Key fingerprint = D4A8 4860 535C ABF8 BA97  25A6 F4F2 1829 9615 02AD
Non-GPG signed mail gets read only if I can find it among the spam.




More information about the users mailing list