On 24 Apr 2002, Stephen Lord wrote:
> On Wed, 2002-04-24 at 14:09, Adam Milazzo wrote:
> > I will try that, but I'd like to note one thing. The drive actually had
> > everything deleted with [essentially] 'rm -rf /mnt/drive/*'. It's not
> > corrupt (just deleted), and I think mounting it might show that it's just
> > here:
> Well, durrrr!! doing too many things at once. Yes, you deleted it all,
> so none of this will help. Sorry about that.
It would be interesting to know if there is anything that could
(at least in principle) be done in this situation.
A while back, we had all the scratch files deleted on a job that
had been running on our cluster for 20+ days. They were on ext2 filesystems,
and I found the wonderful e2undel <http://e2undel.sourceforge.net/> which
almost did what I needed.
It essentially scans the FS for not-in-use inodes with nonzero dtimes,
gives you a list of them, and lets you pull them off. I hacked it up to
look for files with no refs, but zero dtimes (since the app hadn't quit
on most nodes, the files were in limbo) and managed to recover about all
of the 160 GB of data.
I'm not asking if such a tool exists for XFS, mind you, but sooner
or later, given that we have a few TB of XFS in house, I can guess that
I'll want to know if the info to do this still exists on the platter,
and if I could get it through xfsprogs-devel...
Chris J. Bednar
Director, Distributed Computing Product Group