I have a client with a missing files problem.

Around midnight Jan 25/26, TSM did an incremental on a WinNT box.  All three
file systems show as completed in "q fi  f=d".  Some time thereafter, but
before 1/28, the harddrive went south.  On 1/29 directories from c$ were
restored on a new machine.  "Thousands" of ifles are missing.

Where did they go?  "Select" on ll_name for the appropriate hl_name and
node_name from backups finds nothing.  "Select" on hl_name for the appropriate
ll_name and node_name from backups finds nothing.  Combine these queries with
the completed backup from 1/26 and what can we believe but that the files were
not there to backup.  Moreover, the sysadmin says there were file-shares on
another machine(s), which have yet to be identified.

However, the owner says the missing files were on his harddrive before and
after the last sucessful backup.  As he is adamant that the missing files
cannot be on a share somewhere, I have suggested that his hardware problem
crept up on him, making the directory gradually and increasingly unreliable
during backup.  I know this is clutching at straws, but in almost a decade,
this is the first instance of "missing files".

Can anyone suggest a better explanation, or ste of possible explanations, than
I have offered?


Fred Johanson