Add properties as needed to configure multicast cluster member discovery, or well-known address cluster member discovery. Also define the backup count for cluster objects.
If you used the Wizard to set the object management type, a property group called coherence contains all the multicast discovery properties and the backup count property. You can add and delete properties as required for your discovery configuration.Groups are for convenience only. You can group properties differently or not at all, as you find convenient.If you are using well-known addresses for cluster member discovery, delete the multicast properties and instead add the well-known address properties.
The backup count specifies the number of members of the distributed cache service that hold the backup data for each unit of storage in the cache. Recommended values are 0, 1, or 2.Value of 0 means that in the case of abnormal termination, some portion of the data in the cache will be lost. Value of N means that if up to N cluster nodes terminate at once, the cache data will be preserved.A backup count of 1 means one server plus one backup, that is, two cache servers (or storage enabled nodes if cache servers are not used).To maintain the partitioned cache of size M, the total memory usage in the cluster does not depend on the number of cluster nodes and will be in the order of M*(N+1). If you will define cluster members using multicast discovery properties, use the properties in this section, instead of those in the section Well-Known Address Properties. Use this setting if multicast discovery is used and if you need a non-default value. Specifies the multicast IP address that the socket will listen to or publish on.Possible values are addresses between (and including) 224.0.0.0 and 239.255.255.255.Default value is 224.3.3.1 Use this setting if multicast discovery is used and if you need a non-default value. Specifies the port that the socket will listen to or publish on. Specifies the time-to-live setting for the multicast, that is, the maximum number of "hops" a packet can traverse. A hop is defined as a traversal from one network segment to another via a router.For production use, set this value to the lowest integer value that works. Setting the value too high can use unnecessary bandwidth on other LAN segments and can even cause the operating system or network devices to disable multicast traffic.On a single-host cluster, set to zero (0). On a simple switched backbone, set to 1. On an advanced backbone with intelligent switching, it may require a value of 2 or more.Note: A value of 0 is intended to keep packets from leaving the originating machine. However, some operating systems do not implement this correctly, and the packets may in fact be transmitted on the network.
• For multicast discovery when more than one cluster is running on the same subnet (localhost is required but not localport in this case.)
• Default values are provided at the cluster level. However if you need specify these properties at the PU level, add them as Processing Units tab properties and provide the values as needed. As needed, you can set the value of the localhost property to the value localhost. However, if localhost is used as the loop back address (127.0.0.1) you must enter a machine name or IP address.Default value is localhost. Specifies the port that the socket will listen to or publish on.Note If a specified port is not available, the object management layer (by default) increments the port number until it finds an available port. Avoid potential conflicts by choosing a number that is not close to a port used by other software in your environment.Tip To turn off auto-incrementing, add this property: Note If you will discover cluster members using well-known addresses, use the properties in this section, and remove the multicast discovery properties shown in the section Multicast Discovery Properties.Provision is made for up to six well known addresses, using numbers 1–6. If you need more than six, see Overriding and Extending the Operational Deployment Descriptor for information about adding more well-known addresses. However this is seldom needed. The addresses and ports for machines used by the well-known address cluster discovery protocol. At least one of these machines must be running at any time so that others can join the cluster.tangosol.coherence.wka1 10.97.118.151tangosol.coherence.wka2 10.97.118.152Also at the Processing Units tab, configure localhost and localport properties for one processing unit that will be deployed to the WKA machine. Set the localhost value to the value of the wkan property. Set the localport value to the value of the wkan.port property. (See Localhost and Localport Properties)Note You can configure two well-known addresses for the same machine, and use a different port number for each. In this case you would also configure two processing units, each of which matches one set of WKA properties.Tip To turn off auto-incrementing, add this property: If you have customized the cache configuration file, specify the file path and name of the customized file. The default file is located in the BE_HOME/lib/ext/coherence.jar file and is called: An auto-incrementing feature ensures that a different port is used if one specified is already in use. However in various situations you may want to turn off this behavior. For example, if you use TIBCO BusinessEvents Monitoring and Management, and the MM server runs on the same machine as any of the monitored cluster engines, you must explicitly ensure that all ports used by MM and the monitored cluster are unique. Therefore the auto-incrementing feature may not be appropriate.
Copyright © TIBCO Software Inc. All Rights Reserved.