Problem
When loading large sampling results in the z/OS DB2 knowledge base the sampling function uses more and more memory until any limit and stops with out of memory message. This happens when sampling function is performed in Data Express Windows Client (Data Builder) for z/OS.
Resolution
The solution of the problem requests Data Express Windows Client for Distributed Systems to be installed. To avoid the problem with the out of memory message you need to produce the sampling.log file using the sampling function in Data Builder for Distributed Systems and then to load the sampling results using the job ‘’ Load Distributed Sampling Results’’ in Data Builder for z/OS.
The all process includes the following steps:
- 1. Start the Distributed Sampler in Data Builder for Distributed System.
- 2. Select the machine id and company that contains the data stores enabled for sampling as well as the sampling parameters you need.
- 3. Press Export. The sampling.dat file will be generated in the ORACLE/config or ODBC/config directory depending on the Data Store Type with which the data stores are loaded – ORA(OCI) or ODBC.
- 4. Start sampling.exe directly from:
a) The directory …\\Micro Focus\\Data Express 4.0\\ORACLE, if the data stores are loaded using ORA(OCI) connection. After finishing the job the sampling.log file will be produced in the ORACLE/log directory.
b) The directory …\\Micro Focus\\Data Express 4.0\\ODBC, if the data stores are loaded using ODBC connection. After finishing the job the sampling.log file will be produced in the ODBC/log directory.
- 5. Transfer the sampling.log file in a TEXT type format to the sequential file with RECFM FB and LRECL 80 on z/OS machine using the FTP program.
- 6. Start the ‘’Work with Jobs’’ menu of Data Builder for z/OS.
- 7. Select the job ’Load Distributed Sampling Results’. In the Secondary Options menu enter the data set name of this sequential file.
- 8. Run the job.
Incident # 2479567
Old KB# 36132
#MFDS
#Enterprise
#EnterpriseDeveloper
#DataExpress
