Re: Using subfile backup for specific filetypes ?

2004-04-27 18:51:31
Subject: Re: Using subfile backup for specific filetypes ?
From: TSM_User <tsm_user AT YAHOO DOT COM>
Date: Tue, 27 Apr 2004 15:45:52 -0700
The following might also be interesting.
ProblemWhen using adaptive subfile backup, at what percentage change in 
comparison to the base file does the IBM Tivoli Storage Manager Backup-Archive 
client resend the entire base file instead of continuing to create delta files? 
 SolutionAdaptive subfile backup keeps track of its efficiency by evaluating 
the ratio of the size of the previous delta file processed to the size of the 
base file. If the ratio from the last backup exceeds 0.60, the entire base file 
will be resent to the Tivoli Storage Manager server. This ensures that the 
adaptive subfile backup process does not try to create delta files which 
approach or exceed the size of the original base file.
ProblemWhen using adaptive subfile backup, when is a byte-level versus 
block-level subfile algorithm employed?   Solution
The decision whether the adaptive subfile backup is based on a byte-level or 
block-level algorithm is determined by the file size when the base file is sent 
to the Tivoli Storage Manager server. In general, if the file size is greater 
than or equal to 3 MB, a block-level algorithm will be employed. If the file 
size is less 3 MB, a byte-level algorithm will be employed. The exception to 
this rule is when the adaptive subfile backup cache is relatively small, that 
is less then 9 MB. In this case, if the file size is greater then 1/3 of the 
entire cache size, a block-level algorithm will be employed, otherwise a 
byte-level algorithm will be employed.

ProblemWhat are the file size limits used to determine if adaptive subfile 
backup will be used?   SolutionAdaptive subfile backup is not used for very 
small files (less then 1KB in size) or for very large files (bigger then 2GB in 

All of the above can be found by searching at  for 
subfile backup.

"Prather, Wanda" <Wanda.Prather AT JHUAPL DOT EDU> wrote:
TSm compares each filename on the drive to the dsm.opt file, going from the
As soon as it finds a match, it takes action accordingly for that file and
does no further checking.

So, I believe your syntax is correct.
It will

- backup up only files in the MyFolder directory and its subdirectories on
the D: drive, nothing else
- backup any files ending in .msx using subfile
- backup any OTHER files not using subfile

Remember that
1) You must also turn on SUBFILE on the server side (SET SUBFILE CLIENT)
2) The FIRST backup you do this way will still push all the data, since your
first subfile backup has to create a BASE file.
Starting the SECOND backup you do this way, you should see less data because
you will only get DELTA files created.

-----Original Message-----
From: Brian Ipsen [mailto:bin AT PROGRESSIVE DOT DK]
Sent: Sunday, April 25, 2004 12:02 PM
Subject: Using subfile backup for specific filetypes ?


I have a question about the client dsm.opt file...

If I want to make backup of a directory-structure (all files) in a "normal"
manner, I just use the INCLUDE statement. But what is I want all files
backup up using incremental - but files with the extension .MSX in that
folder structure should be backed up using subfile option ? What would this
look like in the dsm.opt file ?

For the normal backup I have:

Domain D:
Exclude *
Include d:\MyFolder\...\*

But in this structure there might be


etc.. These files are quite large and need to be backed up over a line with
limited bandwidth, and that is why I want to use the subfile option...Would
the dsm.opt then look like this:

Domain D:
SubFileBackup Yes
Exclude *
Include d:\MyFolder\...\*
Include.subfile d:\MyFolder\...\*.msx

- or are they listed in the wrong order ??



Do you Yahoo!?
Win a $20,000 Career Makeover at Yahoo! HotJobs