Attention Flash 10280
TSM servers configured to use simultaneous write may leave orphaned
rows in database tables.
Problem: TSM servers configured to use simultaneous write may leave
orphaned rows in database tables. This is caused by a transaction being
aborted. Transactions may be aborted as a result of an error during
processing. Transactions may also be aborted normally as part of prompting
a client to see if the server should wait for the mount of a volume. These
orphaned database rows may cause a variety of operational problems for the
server depending upon the database table affected. Typically, the orphaned
rows caused by this problem were in the AS.Segments or the DS.Segments
server tables and messages such as the following are issued:
ANR01012E Error 1 inserting row in table "AS.Segments".
ANR0102E Error 1 inserting row in table "DS.Segments".
Who is Affected: TSM servers configured to store data to a primary storage
pool and one or more copy storage pools simultaneously.
Recommendation: This problem has been fixed by APAR IC36975 which is
included in levels 126.96.36.199 and 188.8.131.52. Until you can install the fixing
level you can do the following to prevent the problem. Disable
simultaneous write operations by updating the storage pool definition for
the primary storage pool to not have any copy storage pools. This can be
done by issuing the following command:
UPDATE STGPOOL <stgpoolname> COPYSTGPOOL=''
If it is not possible to disable simultaneous write operations, the risk
of encountering this problem can be reduced if the primary storage pool is
a sequential storage device as opposed to a DISK device. So, if the
primary storage pool is DISK, change the storage pool destination to be a
storage pool that uses a sequential media storage device. If the server
configuration is changed so that the primary storage pool is now a
sequential media device, this will require an additional tape drive (mount
point) is available during backup operations. This may require the
MAXNUMMP setting for a given node to be increased using the "UPDATE NODE
xxxx MAXNUMMP=yy" command. It may also require changing when backups are
run to accommodate the available tape drive resources on the server
relative to the number of concurrent backup sessions being run using
Conclusion: The TSM server has been corrected to prevent orphan rows from
being left in database tables during simultaneous write operations where a
transaction is aborted. APAR IC37275 was also opened for this problem. In
the cases where the extraneous row was in the DS.Segments table or the
AS.Segments table, the REPAIR STGVOL command has been enhanced to correct
this situation. The fix for APAR IC37275 is available in the server levels
184.108.40.206 and 220.127.116.11.
View this flash 3 ways :
go to :
search on 10280
If the above information represents information that you would like to
receive automatically, please consider registering for mysupport. With
mysupport, you can have flashes/patch info sent to your email automatically
for quick viewing/evaluating once you subscribe.
Sam J. Giallanza
Tivoli Certified Consultant
Field Issues Manager
Field Input Communications (FIC)
giallanz AT us.ibm DOT com
520.799.5512 - T/L 321.5512
Our new web Support site and KB is at :