I had a somehow good response from Pavel:
"Just hit reset before bdflush writes it to disk!"
So, wouldn't this be an ideal place for this? If we have an update
interval of 60 seconds, then we place every 'delete' into the second
update from now. So we give a minimum recovery-time of 60 seconds up to
119 seconds.
Still, this wouldn't fix the problem when a user comes to the
administrator and says 'I just deleted the wrong file. Can I have it
back?'
If a filesystem is full, then it is full and you cannot undelete any file.
BUT in most 'real life companies' (not on my hd) there is always space
left on the HD, i.e. the admin panics if space drops below 20%. The reason
for this is, a full filesystem is ""read-only"" and noone can continue his
work.
Indeed, this creates need for a 'de-fragmentation-tool' for Linux. I tried
defrag-0.6, but it simply crashed (I know that ext2 doesn't want to
fragment...)
Kurt
----------------------------------------------------------------
Let me make myself perfectly clear: Step 1: Find plan!
Step 2: Save world!
Let's get crackin'... Step 3: Get outta my house!