31101 - Database replication to slave failure

Alarm Group:

REPL

Description:

Database replication to a slave database has failed. This alarm is generated when:

• The replication master finds the replication link is disconnected from the slave.
• The replication master's link to the replication slave is OOS, or the replication master cannot get the slave's correct HA state because of a failure to communicate.
• The replication mode is relayed in a cluster and either:
  • No nodes are active in cluster, or
  • None of the nodes in cluster are getting replication data.

Severity:

Critical

Instance:

May include AlarmLocation, AlarmId, AlarmState, AlarmSeverity, and bindVarNamesValueStr

HA Score:

Normal

Auto Clear Seconds:

300

OID:

comcolDbRepToSlaveFailureNotify

Cause:

Alarm 31101 raises when:

• The replication master finds the replication link is disconnected from the slave.
• The replication master's link to the replication slave is OOS, or the replication master could not get the slave's correct HA state as a failure to communicate.
• The replication mode is relayed in a cluster and either:
  • No nodes are active in cluster, or
  • None of the nodes in cluster are getting replication data.

Diagnostic Information:

1. Verify the path for all services on a node:
   1. In a command interface, type path.test -a <toNode> to test the paths for all services.
2. In a command interface, use the path test commands to test the communication between nodes:
   1. Run the command, iqt -pE NodeInfo to get the node ID
   2. Then, run the command, path.test -a <nodeid> to test the paths for all services
3. Examine the Platform savelogs on all MPs, SO, and NO:
   1. Run the command, sudo /usr/TKLC/plat/sbin/savelogs_plat
   2. The plat savelogs in the /tmp directory.

Recovery:

1. Verify the path for all services on a node by typing path.test –a <toNode> in a command interface to test the paths for all services.
2. Use the path test command to test the communication between nodes by typing iqt -pE NodeInfo to get the node ID. Then type path.test -a <nodeid> to test the paths for all services.
3. Examine the Platform savelogs on all MPs, SO, and NO by typing sudo /usr/TKLC/plat/sbin/savelogs_plat in the command interface. The plat savelogs are in the /tmp directory.
4. Check network connectivity between the affected servers.
5. If there are no issues with network connectivity, contact My Oracle Support.