e2fscking problem after using ext2ifs
 

lo all,

i have been using ext2ifs for 10 months, and after 500gb drive got full i got few I/O errors.
i booted linux, unmounted, and fsck'ed. after 4-5 days of cloning multiple blocks i got to
optimizing dirs and got segfault. after that i stupidly run fsck with -y option and it deleted my journal
and superblock :D after that i used differend available superblocks (-b x), and drive got more corruputed...
errors are currently so numerous that looking errors and pressing yes to questions would take many many hours.
fsck with -p option usually fails to cloning multiple blocks or error reading block x (Attempt to read block from
filesystem resulted in short read) while reading indirect blocks of inode y.

now i agree that fscking can be dangerous, for my mood & drive. my plans are buying another 500gb drive for raw backup
and searching/learning few weeks about data-backup, journal-backup and fscking. is this really necessary?

any good links would be nice 2 have.
thanks.

t: teh_stupid1

p.s some info
# uname -a
Linux c-laptop 2.6.22-14-generic #1 SMP Tue Feb 12 02:46:46 UTC 2008 x86_64 GNU/Linux
# fdisk -l | grep sda4
/dev/sda4 974 60801 480568410 83 Linux
# dmesg | tail -n 3
[23871.163659] attempt to access beyond end of device
[23871.163668] sda: rw=0, want=753308173, limit=690022279
[23871.163674] Buffer I/O error on device sda4, logical block 184419231
# lsusb | grep West
Bus 003 Device 002: ID 1058:1001 Western Digital Technologies, Inc.

Yes, backup is a basic part of information technology. Here on LinuxQuestions we get a surprising number of posts like yours where people present complicated damage to a file system and ask how to save the data. The answer is that you have gone well past the point where the quickest and easiest solution is to restore from backup.

Here is a good in depth discussion of backup:

http://en.wikipedia.org/wiki/Backup

------------------
Steve Stites

In addition to what Jailbait already wrote here's a different take on things. Not that the conclusion is different though...

Not fscking, but 0) filling the drive to the brim and beyond using 1) not the "default" tools. That is dangerous. Too late now, but the first thing you should have done if the data was valuable was (read, ask around and only then) to mount the disk readonly and make a bit copy. (Because of disk errors that probably would have required dcfldd, ddrescue, dd_rescue or rdd most of which are on the HELIX Live CD.) With header / footer-based carving like Photorec, Foremost and other forensic tools can do you already have limited chances (because of it being a journalled filesystem plus the use of a limited set of headers / footers) and not even a guarantee for recovery but with each operation you performed on the data before making a true copy of the disk you have seriously degraded chances for recovery further.

While what you've written does not allow anyone to have an objective or conclusive fix on the fscked up state of your data, I do seriously doubt you'll be recovering much partial data, let alone items in full. The question "is it necessary" can IMHO only be answered by you: it'll take time learning, it'll take time recovering and even then there's absolutely no guarantee at all.