Arch Linux

louis058

Member

Registered: 2010-11-12

Posts: 31

A 'kind-of' corrupted bzip2 archive

Hey, firstly, I'm a noob to Arch Linux, and heck, I'm pretty much a noob to linux as I started with Ubuntu just a month ago, anyway...

I backed up 4.7gigs of data with fwbackups to cross to Arch a few days ago, and didn't verify it (cos I didn't know about that), and now, it's corrupt somewhere. I've tried using bzip2recover, which has given me a lot of chunks of data (over 5000 blocks), but following the instructions here, I tested all the data, and the test results showed that there wasn't a single corrupt block...

Anyway, I put them all together following another tutorial here, and after trying to unzip the tar archive, it came up with another error.

So currently, I have an archive that is corrupt, but isn't corrupt. I have noticed two things though:

Firstly, a lot of the errors are saying that the file 'ends unexpectedly'.
Secondly, when I try to use head or tail to get a chunk of the big archive, I end up with an archive that is the EXACT SAME SIZE. This implies that head or tail has no effect on it (I tried head with -2000, or the first 2000 lines, and got an archive that's still 4.7 gigs).

I've currently got the latest version of bzip2 and tar.

linux-ka

Member

From: ADL

Registered: 2010-05-07

Posts: 232

Re: A 'kind-of' corrupted bzip2 archive

What have you exactly done with tail and head?

When using bzip2recover, and following the first link you mentioned, how many files could you untar/bunzip. When bzip2 -t doesn't find corrupt parts, it must be possible to unpack these packages.

Further, if you still have the original set of data, just create a backup using

 tar vJpcf backup.tar.xz / --exclude=/dev --exclude=/proc --exclude=/sys

extract the tarball via:

tar xf backup.tar.xz

.

louis058

Member

Registered: 2010-11-12

Posts: 31

Re: A 'kind-of' corrupted bzip2 archive

With tail, I followed the instructions of the second website I mentioned in my first post, where I did something like:

tail -c +17185 before.tar > after.tar

and for the head, I followed the man page (I think), and just kind of did the opposite of the above, with:

head -c -2000 before.tar > after.tar

At least, I think I did that. I can't remember whether it was exactly like that, except I know that I almost certainly did the 'head -c -2000' bit.

With bzip2recover, it produced 5368 blocks, and after following the instructions of both websites (I had multiple attempts at this), I ended up not being to unzip anything. It produced an error every time. I even tried cpio, and that did 'file ends unexpectedly' as well.

I suspect either the archive actually does end abruptedly (maybe the backup process somehow stopped before the end?), or there is a bad header. Either way, my head and tail commands aren't working for me.

louis058

Member

Registered: 2010-11-12

Posts: 31

Re: A 'kind-of' corrupted bzip2 archive

Looks like no one knows what my problem is. I do have another backup of my most important document files, so I guess I'll just leave the archive.