ADSM-L

Re: Urgent question: Version retention

2002-03-05 12:22:48
Subject: Re: Urgent question: Version retention
From: Andrew Raibeck <storman AT US.IBM DOT COM>
Date: Tue, 5 Mar 2002 12:20:03 -0500
If file a file named 'a' is renamed to 'b', there is no way for TSM or any
other application (except perhaps for the application that performed the
rename, but then that depends on the application) to know that 'b' is the
same file as 'a'. It is not a bug in TSM that it doesn't recognize that
file 'b' used to be file 'a'.

If the file is renamed, then the creation date/time becomes the creation
date/time that the file was renamed, and the last access date remains
unchanged. For example, suppose file 'a' has a creation date of 02/10/2002
and a modification date of 02/15/2002. After file 'a' is renamed to 'b',
the creation date is 03/05/2002, but the modification date remains the
same. So the fact that TSM showed the .avb file as having a creation date
later than the modification date is correct, as it is just reflecting the
file system information for that file at the time TSM backed it up.

You can verify this for yourself. On an NTFS file system, find some file
that you can rename, say, junk.txt. Run these commands:

dir junk.txt
dir junk.txt /tc
ren junk.txt test.txt
dir test.txt
dir test.txt /tc

The first dir command will show the modification time for the file; the
second command will show the creation time. After the file is renamed, the
third dir command should show the same modification time as the first dir
command, but the fourth dir command will show the current time as the
creation time; and it will appear to have been created after it was
modified.

As to how to recover the file: if your RETONLY setting is for 365 days,
then unless this rename occurred a year or more ago (which doesn't seem to
be the case), an inactive version of the .exe file should still be
around... though it too may be corrupt.

Regards,

Andy

Andy Raibeck
IBM Software Group
Tivoli Storage Manager Client Development
Internal Notes e-mail: Andrew Raibeck/Tucson/IBM@IBMUS
Internet e-mail: storman AT us.ibm DOT com

The only dumb question is the one that goes unasked.
The command line is your friend.
"Good enough" is the enemy of excellence.




Cory Heikel <CHEIKEL AT PSU DOT EDU>
Sent by: "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
03/05/2002 08:59
Please respond to "ADSM: Dist Stor Manager"


        To:     ADSM-L AT VM.MARIST DOT EDU
        cc:
        Subject:        Re: Urgent question: Version retention



Agreed, my concern is why the retain only did not kick in at the time the
files was renamed to .avb. Is tsm really not recognizing a change in the
file even though its' type changed? I have displayed active and inactive
versions for these files and it is as if the .exe version never existed. I
have also looked at the expire inventory and it shows only 1 file being
expired for the entire server. I know of at least 18 on this particular
server that are in the same situation. So it doesn't seem to have expired
the .exe version of the files.

Cory

>>> DMalbrough AT TTIINC DOT COM 03/05/02 10:47AM >>>
It seems to me that this was a very rare situation! As you stated, TSM
would
not have changed the extension on the file from .exe to .avb, that was the
anti-virus software or virus itself. Since this file did not change in 33
days this was the only file, which was **ACTIVE** since Nov. 29th, 2001.
Thirty-three days passed by WITHOUT the file changing (Dec 31st, 2001).

So, if this file is changed any point after Dec 31st, it will become
INACTIVE & be flagged for expiration after the next incremental backup. If
expiration has already been run (typically it has if it runs daily) the
now
INACTIVE copy hqb.exe will drop off and this file (corrupted) hqb.avb will
be ACTIVE for 33 days (beginning March 2nd,2002) until it is changed!


The only thing that could've gotten you out of this situation was a longer
"Retain Extra Versions"!

Only my thoughts!!!!!

Regards,

Demetrius Malbrough
UNIX/TSM Administrator