| Message Text: |
&6 tasks used for parallel index created over file &1. |
| Cause Text: |
&6 is the average numbers of tasks used for an index created over
member &3 of file &1 in library &2.
If file &1
in library &2 is a logical file, then member &9 of physical file &7
in library &8 is the actual file over which the index is being built.
A file name starting with *QUERY or *N for the file indicates
a temporary result file is being used.
The query optimizer
has calculated that the optimal number of tasks is &5 which was limited
for reason code &4. The definition of reason codes are:
- The *NBRTASKS parameter value was specified for the DEGREE parameter of
the CHGQRYA CL command.
- The optimizer calculated the number of tasks which would use all of the
central processing units (CPU).
- The optimizer calculated the number of tasks which can efficiently run
in this job's share of the memory pool.
- The optimizer calculated the number of tasks which can efficiently run
using the entire memory pool.
The database manager may further limit the number of tasks used for the
parallel index build if either the allocation of the file's data is not evenly
distributed across disk units or the system has too few disk units. |
| Recovery Text: |
To disallow usage of parallel index build, specify *NONE on the query
attribute degree.
A larger number of tasks might further improve
performance. The following actions based on the reason code might allow the
optimizer to calculate a larger number: - Specify a larger number of tasks value for the DEGREE parameter of the CHGQRYA CL
command. Start with a value for number of tasks which is a slightly larger
than &5 to see if a performance improvement is achieved.
- Simplify the query by reducing the number of fields being mapped to the
result buffer or by removing expressions. Also, try specifying a number of
tasks for the DEGREE parameter of the CHGQRYA CL command
as described by reason code 1.
- Specify *MAX for the query attribute degree.
- Increase the size of the memory pool.
|