6.26 Checking the operation when the maximum number of hosts is changed (when using HA Monitor Extension)

If you use HA Monitor Extension to change the maximum number of hosts, you must check whether all packets are transmitted completely via monitoring paths. Perform this checking on all hosts in the hot-standby configuration.

Before performing the procedure shown here, if necessary, install the net-tools package so that the netstat command can be used.

The following figure shows the checking procedure.

Figure 6‒66: Procedure for checking the operation when the maximum number of hosts is changed

To check the operation shown in the figure:

Start HA Monitors on all hosts in the hot-standby configuration.

Use the server and host status display command (monshow -c command) to check whether HA Monitors started successfully.

For details about starting HA Monitors, see (1) Starting HA Monitor.
Start servers on all hosts in the hot-standby configuration to enable hot standby processing.

Use the server and host status display command (monshow command) to check whether the systems are ready for hot standby processing, or check messages to determine whether the systems are ready for hot standby processing.

For details about starting servers, see (3) Starting servers.
Check UDP statistical information for a change in UDP packet reception errors.

Execute the netstat -s -u command to obtain the value for packet receive errors in the UDP statistical information. Repeat this step several times at the host failure monitoring interval (specified in the patrol operand in the HA Monitor environment settings) to check for any upward trend.

If there is no change in the UDP packet reception errors values, packets are being received completely, in which case the operation check is finished.

If the values for UDP packet reception errors are increasing, go to step 4.
Adjust the alive message transmission interval.

Whether all packets are transmitted completely depends on the machine's performance and environment, while the performance required by HA Monitor depends on the volume of alive messages that are transmitted. Therefore, use the alive_interval operand in the HA Monitor environment settings to adjust the alive message transmission interval and the volume of alive messages to be transmitted.

Adjust the alive_interval operand value as follows:

If the alive_interval operand value can be increased

Increase the alive_interval operand value.

If the alive_interval operand value is already set to the maximum

Delete the alive_interval operand from the HA Monitor environment settings.
Check the operation again.

Repeat steps 1 through 4. If there is no change in the value for UDP packet reception errors in step 3, the operation check when the maximum number of hosts is changed is finished.

Important: Specify in the alive_interval operand the same value on all hosts in the hot-standby configuration. If you have changed the alive_interval operand value on one of the hosts, also change the operand's value on all other hosts in that hot-standby configuration.

To Page Top