Veritas NetBackup™ Appliance High Availability Reference Guide
- Overview of NetBackup Appliance high availability
- Deployment of a high availability configuration
- Configuration and performance guidelines
- Disaster recovery
- Appendix A. HA Deployment Checklist
Performance tuning for the node failure
In a complete high availability (HA) configuration, when the NetBackup jobs are running, the node where the virtual host name is running may crash or be turned off. It typically can take more than 25 minutes to invoke the automatic failover function.
You can reduce the response time as follows:
Reducing the interval of the time consuming parameters
- In the Java-based NetBackup Administration Console, expand NetBackup Management > Host Properties > Master Servers in the left pane.
- In the right pane, select the master server.
- On the Actions menu, click Properties.
- In the properties dialog box left pane, select Timeouts.
- Change the value of the Job retry delay to 1.
The expected time interval for the services to be switched over is one minute.
- Change the value of the Schedule backup attempts to 12 tries per 1 hour.
This change increases the retry frequency. In this case, the retry limitation may not cause the retry failure.
- Click Apply to make the changes take effect.
- On the master server, locate the file
- Use a text editor, such as vi, to open the file.
- Change the value of the net.ipv4.tcp_keepalive_time to 60.
- Make the change take effect by executing the following command:
sysctl -p /etc/sysctl.conf
- Verify that the change takes effect by executing the following command:
sysctl -a|grep tcp_keepalive
After these changes take effect, the automatic failover for the node failure only takes three minutes to take effect.