Dont swap HDD's on different byte ordering machines please. Data Loss
Story:
I have a x86 Sun Workstation, which is a little-endian
Platform, and I wanted to take a disk with my data on it and place it in a work Sparc Workstation, a big-endian Platform.
Sound simple right, I'm using UFS on the disk, right! right??!!!
Well, I should of read this bug and used some common sense:
4054544
RFE: wish to support different endian byte order
for Solaris. This is regarding the UFS file system, ANYWAYS in the comments there in this bug report which says:
Code:
This was after using other tools, data carving, etc, to find files. The biggest problem of all, it would find files, but not one would preserve filenames. Who wants to go through 5k images, sorting, etc, etc. I guess I could've wrote some code and tried to find it myself, but I have better things to do. Well, TODAY, this MORNING, I found another
software solution that was only 45...... Perhaps the most difficult problem to solve is that even if you do have a
file system that understands both formats, you still can not yank a disk
off a sparc machine and plug it into the intel system and have it work.
The sd driver has intimate knowledge of the disk label and partition
information, which is totally incompatible between Solaris intel and
Solaris sparc.
UMM, so what do I do, I yank a disk from Solaris intel and place it in
a Solaris sparc station, and proceed to wipe the partition of my data
disk.......and my backup data disk...SMART, (this was only because I replied YES to the format question asking me if I wanted to LABEL the HARD DRIVE). Ha, sometimes its so easy to not even think and just click yes.
So the last month, I've been scrambling to recover all this lost data. Which dates back from 98 probably. 10 years of data.
Well, after trying to manually recreate the partition table and finding
out I wiped out all my superblocks too (Dont run a fsck on a disk and
accept changes), I was about to lay down 500 dollars on this program
that found all my data.
If only took me over a month to find it, but it was worth the 450 dollars saved. Thank you GOOGLE.
So the moral of the story, back your data up, not twice, but three
times in different formats. Looking back I should've followed a couple
rules.
1. Always keep the original media in place and do a COPY to the new machine.
2. backup backup backup (should've had DVD backups)
3. buy a external HDD and keep it around as well.
4. Dont swap HDD's with different architectures, even if its the same OS. Do I really have to say this?
And if you do try to recover your data by hand, do a DD to another drive before trying to recreate the partition table, etc. My backup disk didnt work w/ the file recovery program after this, luckily I had a backup.....Any by the way, services to get back over 100gbs of data costs a LOT of money.
Posted at 10:31PM Jan 13, 2008 by Eric Shobe in Personal | Comments[4]
Today's Page Hits: 23
ZFS???
Posted by 85.178.51.236 on January 14, 2008 at 01:02 AM PST #
Use ZFS and this whole blog entry becomes pointless, because ZFS just does it automaticly...
Posted by James Dickens on January 14, 2008 at 07:57 AM PST #
Of course I should use ZFS. so thats what I did after the crash. But what if, just if, somehow my ZFS volume was corrupted. Not *likely* to happen, at least not replicated my steps in the above post, but maybe....
http://opensolaris.org/jive/thread.jspa?messageID=165495
Posted by Eric Shobe on January 14, 2008 at 09:41 AM PST #
Actually, single bit errors on hard drives are more common than people suppose, and can cause silent data corruption. I'd recommend setting up a ZFS mirror so that ZFS can automatically repair any damage. Or use 3 drives with raidz.
Posted by Mark J Musante on January 16, 2008 at 08:17 PM PST #