IBM Support

What is DB2 Accounting Class 2 Not Accounted Time?

Question & Answer


What is DB2 Accounting Class 2 Not Accounted Time?


The following formula defines DB2® Class 2 Not Accounted Time when no parallelism is involved:

DB2 Class 2 Not Accounted Time = DB2 Class 2 Elapsed time -
(DB2 Class 2 CPU time + DB2 Class 3 suspension time)

Usually, the DB2 Class 2 Not Accounted time is very small or negligible. It represents time that DB2 is unable to account for. However, if you see significant DB2 Class 2 Not Accounted time, it could be the result of any of the following causes (this information might change through future maintenance and DB2 releases as DB2 tries to account for waiting time, without causing performance overhead):

1. Too much detailed online tracing, or problems with some vendor performance monitors. This situation is usually the primary cause of high not-accounted-for time on systems that are not CPU-constrained.

2. Running in a high z/OS paging environment and waiting for storage allocation.

3. Running in a very high CPU utilization environment and waiting for CPU cycles, if DB2 work WLM service class goals are not set properly. When you have HiperDispatch enabled in z/OS, it is more important to set the proper WLM goals and importance to reduce the potential for queuing delays. For more information about HiperDispatch, see the z/OS : Planning Considerations for HiperDispatch Mode white paper.

4. A non-dedicated LPAR losing CP when there are multiple LPARs sharing one processor.

5. If the specialty engines are highly utilized and the SYS1.PARMLIB(IEAOPTxx) member has the settings IIPHONORPRIORITY=NO and IFAHONORPRIORITY=NO.

6. Frequent gathering of dataset statistics (SMF 46 Type 2 records)

7. DD consolidation (z/OS parm DDCONS=YES or/and DETAIL) overhead - APAR II07124

8. Delays for CF Lock requests when the lock structures use system managed duplexing

9. Delays due to asynchronous processing of Lock requests when the CF Lock structure is in a remote CF or requests are converted due to heuristic algorithm.

10. In very I/O intensive environments, the Media Manager might be running out of request blocks. This applies to prefetch read.

11. Waiting for parallel tasks to complete when DB2 query parallelism is used.

12. Waiting for parallel tasks to complete when DB2 utility invokes parallelism.

13. HSM (Hierarchical Storage Manager) dataset recall.

14. DB2 internal retry or delay processing to avoid error conditions such as:

    • Retry process when GBP full condition is encountered. Primary GBP retry wait is in "update commit" wait while the retry/wait from secondary GBP shortage is accumulated under Not Accounted
    • Retry process when write in progress for the pages which belongs to GBP dependent page set. This is typically seen in LOB update/insert process. This is resolved with PM52463 (DB2 9 and 10, Feb. 2012)
    • Retry and delay process in pair wise join when there is only one leg. This should be reduced with PM50592 (DB2 9 and 10, March 2012).
    • Retry and delay process during p-lock negotiation. This Not Accounted time should be reduced significantly with PM48334 (DB2 V8, 9, and 10, Nov. 2011)
15. Wait time prior to the first package allocation. This time is now accounted with PM51653 (DB2 10, Jan. 2012)

16. Format write wait from DB2 utilities.

17. DB2 internal suspend and resume when there are excessive resource contentions as DB2 cannot capture all of suspend/resume process.

18. When the distributed thread acts as the requester, the time waiting for the server in the following cases:

    • If the thread is designated allied with distributed server access, then NOT ACCOUNTED includes the time waiting for the server to respond to the requests if the communications used to the server is SNA/APPC. If TCP/IP communications are used, then this time waiting for the server will be reflected in the class 3 wait, SERVICE TASK SWITCH - OTHER SERVICE. Some of the waiting for server time can be found in the distributed activity section of accounting for the each of the remote downstream distributed servers accessed.
    • If the thread is already a DBAT, then regardless of the communications used to a downstream distributed server, NOT ACCOUNTED will include all the time waiting for the server. Again, some of the server wait time(s) can be found in the distributed activity section of accounting for the known downstream distributed servers accessed.
If none of the above causes apply, you might need to provide the following information to IBM Software Support, to determine the cause of the excessive "Not Accounted" time:
    • DB2 Accounting class 1,2,3,7,8 Statistics class(*) for initial diagnosis.
    • Additional documentation may be necessary, such as detail DB2 performance traces with TDATA(CPU,COR,TRA), or/and dumps from DB2 address spaces with large system trace table, and/or CPU Measurement Facility Samples.

[{"Product":{"code":"SSEPEK","label":"DB2 for z\/OS"},"Business Unit":{"code":"BU053","label":"Cloud & Data Platform"},"Component":"Performance","Platform":[{"code":"PF035","label":"z\/OS"}],"Version":"9.0;10.0","Edition":"","Line of Business":{"code":"LOB10","label":"Data and AI"}}]

Document Information

Modified date:
16 June 2018