Custom Reports Using Data Collectors
Product: Cluster Server
Platform: HP-UX
Product version: 5.1
Product component: Global Cluster
Check category: All
Check category: Availability
Check description: Checks whether the date and time are synchronized across the cluster.
Check procedure:
Check recommendation: It is recommended that the date and time settings are identical on all cluster nodes.
Check category: Availability
Check description: Checks for the presence of VCS application agents in the system.
Check procedure:
Check recommendation: It is recommended that you install the missing VCS application agents listed in the output details.
Check category: Availability
Check description: Checks whether the network interface that is specified in the cluster configuration exists on the system.
Check procedure:
Check recommendation: In the cluster configuration, make sure you specify the correct network device.
Check category: Availability
Check description: Checks whether the route to the IP address exists on the network interface specified in the cluster configuration.
Check procedure:
Check recommendation: On the associated network device, add the route to the specified IP address.
Check category: Availability
Check description: Checks whether the UP flag is set for the network interface specified in the cluster configuration.
Check procedure:
Check recommendation: Make sure that you configure the Device attribute of the NIC resource type to a network interface that is configured on the system with the UP flag set. To set the UP flag on a configured device, use following command:!!Linux:!!# ip link set device_name up!!Solaris/AIX/HP:!!For IPv4:!!# ifconfig device_name inet up!!For IPv6:!!# ifconfig device_name inet6 up.
Check category: Availability
Check description: Checks whether the ToleranceLimit attribute has been set for the VCS NIC resource type.
Check procedure:
Check recommendation: Setting the ToleranceLimit to a non-zero value prevents false failover in the case of a spurious network outage. To set the ToleranceLimit for the NIC resource type, login as root and enter the following command:
# hatype -modify NIC ToleranceLimit n
where n > 0.
Because this command prevents an immediate failover and may compromise the high availability of the affected resource groups, only use this command on transient networks.
Check category: Availability
Check description: The ClusterAddress attribute is a prerequisite for GCO. This check verifies if the ClusterAddress Cluster attribute is set, and it is the same as the virtual IP address in the ClusterService service group.
Check procedure:
Check recommendation: Set the value of the Address attribute of the webip resource to that of the ClusterAddress Cluster atttribute.As root :
1. Get the value of the ClusterAddress Cluster attribute:
# haclus -value ClusterAddress -localclus
2. Modify the Address attribute of the webip resource:
# haconf -makerw
# hares -modify webip Address address
# hares -dump -makero, where address is the output of the first command.
Check category: Availability
Check description: If the ClusterService service group is configured, it verifies that its OnlineRetryLimit is set.
Check procedure:
Check recommendation: Set the OnlineRetryLimit for the ClusterService service group.Enter:
hagrp -modify ClusterService OnlineRetryLimit N
where N >= 1
Check category: Best practices
Check description: Checks whether the IfconfigTwice attribute for the VCS IP resource type is set to 1. Setting the attribute to 1 ensures that when the IP address is brought online or failed over, the system sends multiple Address Resolution Protocol (ARP) packets to the network clients. Sending multiple packets reduces the risk of connection problems after a failover event.
Check procedure:
Check recommendation: Make sure you set the IfconfigTwice attribute has been set to a value of 1 or larger.
Check category: Best practices
Check description: Checks whether the NetworkHosts attribute for the VCS NIC resource type has been configured. This attribute specifies the list of hosts that are pinged to determine if the network is active. If you do not specify this attribute, the agent must rely on the NIC broadcast address. This causes a flood in network traffic.
Check procedure:
Check recommendation: Make sure you configure the NetworkHosts attribute with a list of IP addresses that can be pinged to determine if the network connection is active. To set the NetworkHosts attribute for the NIC resource, login as root and enter the following command: # hares -modify res_name NetworkHosts ip_address where ip_address is a space separated list of IP addresses.