Question & Answer
Question
Object aggregation allows IBM Spectrum Protect (formerly called Tivoli Storage Manager) servers to optimize server internal data movement operations. It is common that servers that store IBM Content Manager objects are not configured for optimized aggregation, so the question is how you can verify IBM Content Manager object aggregation on the backup server?
Cause
The reason for object aggregation being favorable is documented in the technote: Slow disk storage pool backup with small client transactions, http://www.ibm.com/support/docview.wss?uid=swg21246443
Answer
To verify whether your objects are aggregated, you can review the output of the SHOW TRANSFERSTATS administrative command. The following sample shows what output from the command might look like:
show transferstats <STORAGEPOOL>
Statistics for last migration from pool <STORAGEPOOL>
Start date/time: 12/17/15 21:45:27
Elapsed time: 833 seconds
Total wait time: 0 seconds
Number of participating processes: 2
Total duration of all processes: 833 seconds
Total physical files: 45202
Total logical files: 45202
Total bytes: 12334968832
Average logical files per physical file: 1.0
Average physical file size: 266 KB
Number of batch/file transactions ended: 47
Number of batch transactions aborted: 0
Number of file transactions started: 0
Number of file transactions aborted: 0
Statistics for last storage pool backup of <STORAGEPOOL>
Target copy storage pool: TAPE_ARCHIVEPOOL_BSW
Start date/time: 12/18/15 06:02:01
Elapsed time: 81 seconds
Total wait time: 47 seconds
Number of participating processes: 1
Total duration of all processes: 81 seconds
Total physical files: 900
Total logical files: 900
Total bytes: 92741632
Average logical files per physical file: 1.0
Average physical file size: 100 KB
Number of batch/file transactions ended: 1
Number of batch transactions aborted: 0
The sample output shows that no aggregation takes place: The "Average logical files per physical file" fields show the value of 1.0, and each object is smaller than what is defined with TXNBYTELIMITS client option.
Keep in mind that the typical transaction limits for a backup-archive client are defined by the following parameters:
- TXNBYTELIMITS (option on the client): Specifies the number of kilobytes the client program sends to the server before committing the transaction. The range of values is 300 through 34359738368 (32 GB). The default is 25600 KB.
- TXNGROUPMAX (option on the server): The TXNGROUPMAX option specifies the number of objects that are transferred as a group between a client and the server between transaction commit points. The minimum value is 4 objects and the maximum value is 65000 objects. The default value is 4096 objects. The objects transferred are actual files, directories, or both. The server counts each file or directory as one object.
A possible circumvention is the RM_VOLUMES VOL_AGGREGATESIZE parameter:
http://www.ibm.com/support/knowledgecenter/SSAHQR_8.4.3/com.ibm.systemtables.doc/rmvolumes.htm
- VOL_AGGREGATESIZE Maximum size of an aggregate object. A zero (0) value means disabled.
Usually the Content Manager systems run with the default of 0, which seems to be the case with the sample given above.
However, the option is not supported by the Content Manager product for standard IBM Spectrum Protect servers. Read about that restriction in the technote: RM_VOLUMES VOL_AGGREGATESIZE and Tivoli Storage Manager server, http://www.ibm.com/support/docview.wss?uid=swg21699267
Was this topic helpful?
Document Information
Modified date:
17 June 2018
UID
swg21974852