DB2 probe support for cluster failover awareness
search cancel

DB2 probe support for cluster failover awareness

book

Article ID: 408939

calendar_today

Updated On:

Products

DX Unified Infrastructure Management (Nimsoft / UIM)

Issue/Introduction

Does the DB2 probe support monitor DB2 Database 

We want to monitor database connectivity, using the i_check_dbalive with the DB2 probe. However because we have a cluster, one side is currently always reporting the database is down on the standby node.

We need an alarm from the standby node indicating that it has switched from primary to standby during the failover, is this possible? 

Environment

  • DX UIM 23.4.*
  • DB2 probe 4.31

Cause

Currently the DB2 probe does not support cluster awareness.

Resolution

A test build of the DB2 probe exists that provides cluster awareness, the next GA version of the probe (eg. 4.32) will include this new functionality. 

Implementation details are as follows:

  • The probe now saves the role state in a file named hadr_role.state, located in the probe folder.
  • During a failover, the probe uses this state file to determine the role change. Based on this, the Standby DB2 probe will generate an alarm containing the hostname, database, and instance details.
  • Once the system returns to an online state, the alarm will be cleared. Subsequently, the other standby node will raise the same alarms with its respective hostname, database, and instance details.

    Note: For the first failover after deploying this version, no alarm will be raised. This is because the hadr_role.state file does not exist from the previous build, and the probe cannot determine which node was previously primary or standby.

    From the second failover onward, the mechanism will function seamlessly.

For testing this probe open a case with DX UIM Broadcom support. 

Additional Information

Notes:

With the test build that can be provided by support under request:

  • Failover and failback alarms will now be generated at the database level.
  • Each database will maintain its own state file (hadr_role.{instance}.{database}.state)
  • All profiles monitoring the same database will share the corresponding state file.