Incremental Back-up Problem

[LongoMR]

[]

I am currently thinking about possible solutions to a problem
we have come across and would appreciate knowing if anyone else
has had a similar problem and how it was resolved.

Our back up procedure currently consists of a full image-copy
back up once a week (volcopy) with incremental back-ups
taken the other 6 nights. About a month ago we lost a complete file system.
This happened on a thursday (3rd night of incrementals). The restore procedure
consists of restoring the last good full image of the file system and then
overlaying each subsequent incremental on top of that.

The problem this created was due to the fact that between the
time of the last image copy and the time we lost 
the file system several directories
were moved to different names using cpio. When the file system was 
recreated, it first installed the original named directories (which were 
on the volcopied version) then proceded to write the 
files with the new names as they existed on the incrementals. These
were in addition to the original files. The
file system quickly ran out of space and the restore failed. Without 
input from our users we found it impossible to remove the proper
files and directories and complete the restore.

Has anyone else faced this problem? Any thoughts on a possible
solution would be greatly appreciated. I feel that replies to the
net in this situation would generate more insight as solutions are usually
seen and commented upon by many, so please reply to this newsgroup.
As usual, thanks in advance.
	Mark Longo	...!ihnp4!drutx!mrl