A fix is available
APAR status
Closed as program error.
Error description
A performance degradation (in query execution time) is possible when the result of a sparse index is directed to a physical workfile. This can create an excessive amount of GetPages and CPU when sparse index is accessed. When workfile is bigger than the sparse index pool, the sparse index performance suffers from the big workfile access cost. Internally, the cost penalty to the sparse index access needs adjustment when the workfile size is too big. Db2 should skip building the sparse index when the following conditions are met. - The workfile size, 10 times larger than the total sparse index pool size. - The outer cardinality is comparable to the inner table's ONECOMPROWS (DSN_DETCOST_TABLE). The inner table's ONECOMPROWS is larger than 10K. - There exists an index plan on the inner table that covers all sparse index keys and more. When the above conditions are met, the index plan can provide the same or better predicate coverage. Db2 will skip building the sparse index access to avoid huge workfile access cost related to the sparse index access. ADDITIONAL SYMPTOMS: SQLPERFORMANCE PERFM SQLSPARSEINDEX SPARSEINDEX
Local fix
Problem summary
**************************************************************** * USERS AFFECTED: * * All users of Db2 12 for z/OS running * * queries referencing big tables and * * using sparse indexes. * **************************************************************** * PROBLEM DESCRIPTION: * * A performance degradation (in query * * execution time) may occur when the * * following conditions are met. * * (1) The PLAN_TABLE.PRIMARY_ACCESSTYP * * has the value 'T' for a table, meaning * * the sparse index is used and, * * (2) The sparse index size in (1) is * * bigger than the sparse index pool * * size. * * When the conditions above are met, * * the sparse index can spill to * * workfiles. It can cause excessive * * GetPages and CPU. Therefore it will * * lead to suboptimal performance. * **************************************************************** * RECOMMENDATION: * * Apply corrective PTF when available * **************************************************************** A performance degradation (in query execution time) may occur when the following conditions are met. (1) The PLAN_TABLE.PRIMARY_ACCESSTYP has the value 'T' for a table, meaning the sparse index is used and, (2) The sparse index size in (1) is bigger than the sparse index pool size. When the conditions above are met, the sparse index can spill to workfiles. It can cause excessive GetPages and CPU. Therefore it will lead to suboptimal performance. SQLSPARSEINDEX SPARSEINDEX SQLACCESSPATH SQLPERFORMANCE
Problem conclusion
PH45369 will solve the sparse index performance problem described above by detecting the following conditions and skip using the sparse index. (1) The sparse index size is bigger than the sparse index pool and, (2) The inner table size estimate after applying all the local predicates is bigger than 10K and, (3) There exists an index that can cover all the spasre index keys. Please refer to PH46618 for the v13 fix.
Temporary fix
Comments
APAR Information
APAR number
PH45369
Reported component name
DB2 OS/390 & Z/
Reported component ID
5740XYR00
Reported release
C10
Status
CLOSED PER
PE
NoPE
HIPER
NoHIPER
Special Attention
NoSpecatt / Xsystem
Submitted date
2022-04-01
Closed date
2022-06-15
Last modified date
2022-09-01
APAR is sysrouted FROM one or more of the following:
APAR is sysrouted TO one or more of the following:
PH46618 UI81024
Modules/Macros
DSNXOCSC
Fix information
Fixed component name
DB2 OS/390 & Z/
Fixed component ID
5740XYR00
Applicable component levels
RC10 PSY UI81024
UP22/08/06 P F208
Fix is available
Select the PTF appropriate for your component level. You will be required to sign in. Distribution on physical media is not available in all countries.
[{"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Product":{"code":"SSEPEK","label":"DB2 for z\/OS"},"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"12.0","Line of Business":{"code":"LOB10","label":"Data and AI"}}]
Document Information
Modified date:
01 September 2022