5.5.3 Monitoring path configuration

(1) How to connect monitoring paths

You must provide a dedicated TCP/IP LAN for a monitoring path. A TCP/IP LAN can be used in any hot-standby configuration, but it requires a dedicated IP address and port number.

A monitoring path connects two processors on the hosts (in the active and standby systems) that monitor each other. Configure the active and standby systems and the monitoring path that connects them in the same network. If one active system has multiple standby systems, connect the active system and all the standby systems with a monitoring path and configure them in the same network.

To Page Top

(2) Using multiple monitoring paths

If there is only one monitoring path and a failure occurs on the monitoring path, HA Monitor might treat the failure as a host failure and perform the hot standby operation. To prevent hot-standby switchover from occurring at a single point of failure, always multiplex the monitoring path. Note that at least one of the multiple monitoring paths must be used for only monitoring purposes. If there is only one monitoring path, HA Monitor might not be able to determine whether the hot standby operation can be continued.

If the system being used is BladeSymphony, the number of supported LAN adapters is limited. Therefore, before you provide multiple monitoring paths, you must consider sharing a monitoring path with a business application LAN or with the reset path.

If you use SCSI reservation for shared disk, we recommend that you share a monitoring path with a business application LAN. If multiple failures occur on multiple monitoring paths (disabling communication), and the hot standby operation is performed, alias IP addresses might remain assigned to both hosts before the source HA Monitor detects a loss of reservation. If you share a monitoring path with a business application LAN, the source host uses its alias IP address for communication less frequently.

You can use a maximum of six monitoring paths. If you provide multiple monitoring paths and a failure occurs on one of them, HA Monitor can continue monitoring hosts by using another path. Also, you can choose either of the following operation modes:

• Reducing the amount of data load per monitoring path

  The amount of data load per monitoring path can be reduced because HA Monitor uses a different monitoring path for each instance of communication between hosts. If you do not specify a preferred monitoring path, this operation mode is used.

• Specifying a preferred monitoring path

  If you specify a preferred monitoring path in the lan operand in the HA Monitor environment settings, HA Monitor uses the specified monitoring path first for communication between hosts. If a failure occurs on that monitoring path, HA Monitor uses another monitoring path to continue communication.

  Note that if you specify an alive message transmission interval in the alive_interval operand in the HA Monitor environment settings, HA Monitor sends alive messages to all monitoring paths. For this reason, a preferred monitoring path cannot be specified.

If you use redundant LAN adapters by using a kernel extension package in Linux, such as a bonding package and VMware ESXi functions, you can prevent the hot standby operation resulting from a single point of failure without having to provide multiple monitoring paths. In such a case, you must specify a value that is larger than the LAN adapter switching time in the patrol operand in the HA Monitor environment settings because it takes time to switch LAN adapters.

To Page Top

(3) IP addresses to be assigned to multiple monitoring paths

A monitoring path requires a dedicated IP address and port number. In order to provide multiple monitoring paths, you must note the following:

• Use a different LAN segment for each monitoring path.

• Use a different network address for each monitoring path.

A network address consists of a network number and a subnet number.

1. When assigning a different network number to each monitoring path

   There is no limitation to the supported IP addresses. The following table shows examples of IP address combinations.

   Table 5‒5: Examples of IP address combinations (when assigning a different network number to each monitoring path)

   LAN	IP address	Netmask	Network number	Subnet number	Host number
   LAN1	100.2.1.130	255.255.255.0	100	2.1	130
   LAN2	101.2.1.130	255.0.0.0	101	None	2.1.130

2. When assigning the same network number to all monitoring paths and a different subnet to each monitoring path

   You must specify the same netmask.

   In a network configuration where the same network number is assigned to all monitoring paths and a different subnet is assigned to each monitoring path, make sure that you do not specify different netmasks. The example shown below cannot use multiple monitoring paths because its IP address combination has the same network number but different netmasks.

   Table 5‒6: Example of IP address combinations that cannot use multiple monitoring paths

   LAN	IP address	Netmask	Network number	Subnet number	Host number
   LAN1	100.2.1.130	255.255.255.0	100	2.1	130
   LAN2	100.2.2.130	255.255.0.0	100	2	2.130

   The IP address combination shown below can use multiple monitoring paths because it uses the same network number and netmask.

   Table 5‒7: Example of IP address combinations that can use multiple monitoring paths

   LAN	IP address	Netmask	Network number	Subnet number	Host number
   LAN1	100.2.1.130	255.255.255.0	100	2.1	130
   LAN2	100.2.2.130	255.255.255.0	100	2.2	130

   The following figure shows a network configuration using multiple monitoring paths in which the same network number is assigned to all monitoring paths and a different subnet is assigned to each monitoring path.

   Figure 5‒7: Network configuration using multiple monitoring paths

To Page Top