Host or member issues

When you encounter a problem, it is important to diagnose and then troubleshoot the problem. Details are given on how to identify alerts or errors and to diagnose and troubleshoot symptoms that might occur in a Db2® pureScale® environment.

In a Db2 pureScale environment, the db2instance -list command can be run from any host where a member or cluster caching facility is located to get information about problem symptoms that might occur.

The three distinct state elements: member, cluster caching facility, and host can show symptoms such as alerts, errors, and WAITING_FOR_FAILBACK.

If the database is unavailable, reporting of some data is still possible if the Db2 pureScale instance manager is still available on the host because you can use the db2instance -list command if there is no current database connection or if the instance is stopped.

Further basic diagnostic steps include:
  1. Running the db2cluster (-cm) -list -alert command to check for alerts
  2. Checking the db2diag log file for data diagnostic information and error messages
  3. Checking the DIAGPATH directory for any diagnostic files that are produced and FODC directories
  4. Checking the CF_DIAGPATH directory for any cluster caching facility diagnostic data produced
  5. Running the errpt -a command to view the system error log. Look for related log entries at the time of the failure.

You can monitor the state of hosts members and cluster caching facilities.