You must install, provision, and configure a Db2 or Db2 Warehouse instance on Cloud Pak for Data for each Db2 Data
Gate instance.
About this task
A Db2 Data
Gate instance requires either a
Db2 or Db2 Warehouse instance depending on the type
of target database you select when creating a Db2 Data
Gate instance.
- If you plan on using Db2 Data
Gate for
transactional workloads, install and configure Db2 as described in Installing Db2.
- If you plan on using Db2 Data
Gate for
analytical workloads, install and configure Db2 Warehouse as described in Installing Db2 Warehouse.
In either case, the following requirements and recommendations should be followed during
Db2 installation and configuration to ensure
correct operation and optimal performance of Db2 Data
Gate:
Procedure
- Recommended: For the best Db2 Data
Gate performance, create a Db2 or Db2 Warehouse service instance on a
dedicated node with its own dedicated storage and set the storage volume type to
hostPath. The
hostPath storage volume type offers a
lower I/O latency. Always aim for an extremely low synchronization latency and a high load
throughput.
Note that each database instance needs a different dedicated hostPath
location on the host machine.
For more information, see:
- Recommended: For the best Db2 Data
Gate performance, after the Db2 or Db2 Warehouse instance is created, consider
to increase the total size of the Db2 or
Db2 Warehouse log file. Otherwise,
the loading of tables might fail because the log file is too small.
- For test and development systems, consider to increase the total size to about 3 GB by setting
the value of the logfilsiz parameter to
50000 and the value of
the logprimary parameter to 15.
- For production systems, your Db2
administrator should adjust the settings based on the system capacity.
- To avoid losing your logfilsiz and logprimary settings
after an upgrade, change the configuration of the target database as follows:
- As the user
db2inst1, access the container of the Db2 or Db2 Warehouse target database and run a
command to update the configuration. To this end, replace ${DB2U_POD_NAME} with the
name of the server pod name and ${PROJECT_CPD_INST_OPERANDS} with your project name
in the following sample command. Also replace xxx with appropriate values. Then run
the
command.oc exec -it ${DB2U_POD_NAME} -n ${PROJECT_CPD_INST_OPERANDS} su - db2inst1
db2 update db cfg for bludb using logfilsiz xxx logprimary xxx
- Update the database configuration parameters of your cluster:
- Determine the
Db2ucluster CR name of your target database instance and edit it.
To this end, replace ${PROJECT_CPD_INST_OPERANDS} with your project name in the
following sample command. Then run the
command.oc get db2ucluster -n ${PROJECT_CPD_INST_OPERANDS}
DB2UCLUSTER_ID=<db2ucluster_cr_name>
oc edit db2ucluster ${DB2UCLUSTER_ID} -n ${PROJECT_CPD_INST_OPERANDS}
- The database configuration (YAML file) is now in edit mode. Database configuration parameters
are added under
spec.environment.databases.dbConfig. Navigate to this section and
add or change the logfilsiz and logprimary values as shown
in the following example:spec:
environment:
databases:
name: BLUDB
dbConfig:
logprimary: "10"
logfilsiz: "250000"
- Save and exit the YAML file. Then exit the container.
For more information, see:
- Recommended: To further optimize the performance after the creation of the
Db2 instance, consider increasing the values
of the locklist and maxlock parameters. This is to avoid
that too many row locks are replaced with a lock of the entire table while tables are being
synchronized (a process called lock escalation).
- For test and development systems, keep the default value
Automatic.
- For production systems, change the value of the locklist parameter to
500000 and the value of the maxlocks parameter to
13.
For more information, see: