Db2 change selector does not start - Connect_CDC - connect_cdc_mimix_share - 6.x

If the Change Selector does not start from a monitoring and control tool, such as Connect CDC MonCon, check the following:

• Exiting abnormally (for example, the machine crashes) may cause the following problems:

• Connect CDC MonCon may report that the Change Selector is running when it is not.

This is due to the fact that the Change Selector was unable to post a Stopped status in the Connect CDC RP_INFO table. You can check this with the following query:

SELECT from RP_INFO WHERE RP_I_TYPE=’STATUS’

The RP_INFO column indicates the last current status.

From a shell, you can determine whether the Change Selector processes are running by issuing the following command:

ps -er|grep udbcs

This should display the following processes:

|udbcs.sh   <--shell that starts the Change Selector

|udbcs   <--Change Selector parent process

• Error loading file mentions .PID file when attempting to start the Change Selector:

INFO: Writing lock file ‘<name>cts.PID’ for PID <name>
            CRITICAL: Error loading file ‘<name>cts.PID’ Resource temporarily unavailable
            CRITICAL: lock file previously locked by PID<name>

Determine the status of the Change Selector:

• Check the status in Connect CDC MonCon.

• Check the status using command ps -er|grep udbcs.

If you determine that the Change Selector is not running, then there may have been some outage (USS crash, or other) whereby an erroneous Change Selector PID file was left behind in the kernel directory. This file is named:

<model>.<server>_rpzoscs.cfg.PID

The Change Selector creates this file when it is initialized. The file is used to prevent another Change Selector from starting and processing the same model. This could happen, for example, if the Change Selector is already started and you inadvertently attempt to start it again.

If the Change Selector is NOT active and does not start because of this error, then you can delete this PID file.