[NTLUG:Discuss] Data corruption

[terry]

I would try some diagnostics on the  drives
smartct -a /dev/hd?  |less
hdparm -i /dev/hd?
(smartct most importantly)
You might also be interested in hdsentinel
http://www.hdsentinel.com/hdslin.php

On Sat, Jan 23, 2010 at 1:29 AM, LEROY TENNISON
<leroy_tennison at prodigy.net> wrote:
> Borrowing from another thread: "things go wrong".  I think I finally know what hardware is failing after too much data corruption.  Booting from two different hard drives (SuSE 11.2 on a Samsung, Debian Lenny 5.0.3 on a Seagate - used at different times on the same motherboard) I would be working and some event would occur at which point the OS would start acting weird.  As best as I can tell the hard drive simply "went away".  Only solution was to reboot (or try, it never worked).  If the hard disk still had partitions at all they were massively corrupted.  The disks do still work (SuSE was able to find lost partitions and rebuild, e2fsck did repair some of the Debian partitions I tried).
>
> I now have lots of questions.
>
> Given the symptoms, what of the hardware should I most suspect: CPU, RAM, motherboard, power supply?  If I can salvage anything I'd like to but not at the cost of losing a few weeks worth of data which is where I may be right now.
>
> Next question, what's the best possible hope for recovering my data off of the Debian ext3 partition?  e2fsck can't find a superblock and stops there.  SuSE's repair (I know, different distro but both support ext3) says it doesn't have a valid filesystem.
>
> Final question, other than doing full backups frequently (which I don't tend to do on my personal computer which is what this is), are there other disaster recovery options like backing up the superblock or some other technique?
>
> Thanks for any and all help.
> _______________________________________________
> http://www.ntlug.org/mailman/listinfo/discuss
>



-- 
<><