Too Much Oracle Data Stored

[GregE]

I have a problem with some Oracle databases where I'm noticing there is far too much data stored in TSM.

Example. DB1 is 560GB. It's a development database. We back up once a week and we keep two copies, as seen in RMAN as:

RMAN> show all;
RMAN configuration parameters are:
[b]CONFIGURE RETENTION POLICY TO REDUNDANCY 2;[/b]

In TSM, occupancy is 8.8 TB (4.4 onsite, 4.4 offsite). 4.4 TB is far too much data for a 560GB database that has only two versions retained.

After each backup, this is run:
CROSSCHECK BACKUP;
DELETE EXPIRED BACKUP;
report obsolete;
delete obsolete;
....and those commands are working, as I see RMAN activity logged as they delete data.

On the node definition, "BACKDELETE" is "YES."

Is there anything else I should be looking into?

 

[ChrisRees]

Hi

I'm guessing that your RMAN catalog and TSM are out of sync. Have you run tdposync?

http://publib.boulder.ibm.com/infoc....jsp?topic=/com.ibm.itsmfd.doc/ab5u000364.htm
cheers

 

[GregE]

I forgot to mention that I dd run that. It didn't find any discrepancies.....

The TSM Server is Synchronized with the Oracle Catalog or there
are no matching files in the requested filespace.

 

[GregE]

Found 66 of these, which I just deleted....

RMAN retention policy will be applied to the command
RMAN retention policy is set to redundancy 2
Report of obsolete backups and copies
Type Key Completion Time Filename/Handle
-------------------- ------ ------------------ --------------------
Datafile Copy 6831611 Mar 01 2009 14:02:30 /oracle/data/db1/db1.dbf
Datafile Copy 6831612 Mar 01 2009 14:02:30 /oracle/data/db1/dbindx.dbf
Datafile Copy 6831613 Mar 01 2009 14:02:30 /oracle/data/db1/users.dbf
...etc....etc

Ran this.......
CROSSCHECK COPY;
DELETE EXPIRED COPY;

TSM expiration will run later and I'll see if I get a big drop in occupancy.

 

[GregE]

Huge drop!

 

[ChrisRees]

Hi

I wonder how those copies got created in the first place?
I presume they must have been created outside of the normal rman backup routine..

Cheers

 

[GregE]

Well, our Oracle DBA implemented Oracle ASM in the last few months. For at least two databases (the only ones I can find record of the work), the "Datafile Copy"s listed are the exact day of that implementation, so I'm thinking that had something to do with it. Odd part is the growth. I can understand that happening, since the filesystem that used to house the database was deleted, in favor of Oracle ASM, so in my previous post, that path you see is the former filesystem. I'm puzzled by the amount of growth. For the particular database I found the date for, we dropped from 7500GB stored, to 925GB, so it was much more than just one copy of the database that wasn't being deleted.

 

[GregE]

I'm just curious if anyone else has seen or knows about what takes place in TSM's view when an Oracle database is converted to ASM. I don't quite have a handle on why our data continued to grow so much after conversion. I can kind of understand TSM and Oracle TDP leaving expired data at conversion, but for it to continue to grow so much, I don't understand. Anyone have any experiences?