When receiving persistent DB03135W messages, use the following steps to determine if the condition is expected workload lag or a potential task blockage:
- Issue the STATUS_CDC Console Command:
- Run STATUS_CDC to see the current state of CDCL and CDCU tasks.
- Observe the "LAST CDC" vs. "LAST LXX" timestamps.
- Check the "STATUS" column. A status of WAIT CDCL POLL or WAIT CDCU may indicate the task is waiting for work or potentially stuck.
- Analyze MUF Logs for External Blockages:
- Search MUF logs for errors occurring concurrently with the lag.
- Look for DB01709I (Dynamic Extend) messages. If the CDC database is reaching the z/OS maximum number of extents (16), it may cause I/O delays or failures that freeze the CDC process.
- Check for DB03163I messages. A high "WRAP" count in this message indicates the startup option CDC buffer may be too small.
- Check for DB00321I LOG LXX WRAP messages. This indicates the LXX should be much larger to prevent a 'GAP' with CDC that will require manual intervention and substantial complexity to remove.
If STATUS_CDC confirms that the tasks are not making progress, or if logs show resource constraints:
- Priority Check: Ensure the CDCU job is running with sufficient WLM priority. If it cannot keep up with the volume of work, it may require higher importance.
- Dataset Capacity: If the CDCU is running as fast as possible but cannot keep current, the CDC base needs to be large enough to ensure it does not fill while waiting on the CDCU task.
- Buffer Tuning: If DB03163I shows high WRAP values, consider increasing the CDC startup option buffer size (up to 16M).
- LXX Size: If DB00321I WRAP messages are present, increase the size of the LXX.
- Transient vs. Permanent: If the task is running but falling behind due to heavy transaction volume, this is expected behavior. If the task shows no movement in timestamps (e.g., stuck for hours), restart the CDCL/CDCU task to verify if it resumes processing.