WebSphere Application Server Network Deployment, Version 6.1
             Operating Systems: AIX, HP-UX, i5/OS, Linux, Solaris, Windows, z/OS

             Personalize the table of contents and search results

Core group protocol versions

Core group members interact with each other through a variety of protocols such as the discovery protocol, the failure detection protocol, and the view synchrony protocol. Each of these protocols define a set of formatted messages that core group members exchange according to a common algorithm.

A core group protocol consists of a set of formatted messages that the core group members exchange according to a common algorithm. A new version of a core group protocol is provided whenever high availability enhancements require format changes that make the new message set incompatible with the previous version of the message set. The high availability manager can use any supported core group protocol version as long as all members of a particular core group are configured to use the same version.

New protocol versions are added to the product if new messages, or new algorithms are required to support new product features, or to improve core group performance. Because the new messages, or new algorithm might not be compatible with the existing messages or algorithm, a new protocol might not be able to interoperate with the old version of the protocol.

There are two major categories or groups of protocols.

The protocol version settings for each of these two categories are independent of each other.

Whenever a new protocol version is added to the product:

When to select a new core group protocol version

Core group protocol versions are always cumulative. Any functional enhancement that is provided in a previous protocol, is included in any subsequent protocols. It is recommended that you always use the latest protocol version whenever possible. But, before configuring the members of a core group to use a new protocol version, you must make sure that all of the core group members are running at a WebSphere Application Server code level (VRM) that is equal to or greater than the new protocol version.

When you have determined that you can use a newer core group protocol level with a particular core group, use either the IBM_CS_WIRE_FORMAT_VERSION or the IBM_CS_HAM_PROTOCOL_VERSION core group custom property to configure all of the core group members to run with that newer version. You can change the value of this property while the core group members continue to run. After you save and synchronize the changed value to all of the nodes containing core group processes, the high availability manager automatically detects the configuration change and starts using the new core group protocol version to communicate with the core group members.

Avoid trouble: You do not have to restart the core group members when you move from an older protocol version to a newer protocol version. However, you must restart all of the core group members if you move from a newer protocol version to an older protocol version.gotcha

Using the high availability manager protocol to establish transparent bridge failover support

Core group bridges provide the mechanism that is used to represent and manage the cross core group state that is used by WebSphere Application Server components. Part of the management process for this cross-core group state is to perform core group bridge state rebuilds whenever there is a change in the number of running core group bridges in a topology. The core group bridge state rebuild is the means by which core group bridges calculate the ownership, and distribution of the cross-core group state among the running set of bridges.

During core group bridge state rebuilds, cross-core group state can be moved between running bridges. This situation might cause the data to be temporarily unavailable until the bridge has completed the rebuild process. Common symptoms of this problem include:
  • JNDI lookups failing.
  • A WebSphere proxy server, or on demand router generates a 503 response code after a core group bridge failover occurs
  • The following array index out-of-bounds exception occurs:
    [7/9/08 17:12:20:749 EDT] 00000030 UserCallbacks E 
    HMGR0142E: An error occurred in a component called back by the High Availability Manager 
    The exception is java.lang.ArrayIndexOutOfBoundsException at
    com.ibm.ws.cluster.propagation.bulletinboard.BBDescriptionManager.getOrderedBytes(BBDescriptionManager.java:618) 
Best practice: If you are running on Version 6.1.0.19 or later, set the IBM_CS_HAM_PROTOCOL_VERSION core group custom property to 6.0.2.31 for all of your core groups to avoid a possible high availability state outage during core group bridge failover. When this custom property is set to 6.0.2.31, the remaining bridges recover the high availability state of the failed bridge without the data being unavailable in the local core group.bprac
Avoid trouble:
  • If a core group is at a service level that does not support the 6.0.2.31 high availability manager protocol, that core group can still communicate with core groups that use the 6.0.2.31 high availability manager protocol. However, the core group is at a service level that does not support the 6.0.2.31 high availability manager protocol might still experience missing data during core group bridge failover.
  • Transparent bridge failover is designed to hold state data constant during core group bridge rebuilds along the state data path, which is the path that consists of the state provider, one core group bridge in each respective core group, and a state data consumer. Failure scenarios involving core groups with no remaining active bridges might still result in temporary state outages.
gotcha

Determining which protocol version to use

Best practice: It is recommended that you always use the newest protocol version whenever possible. This practice is especially critical for large topologies because most of the recent protocol changes include scalability improvements. However, before configuring the members of a core group to use a new protocol version, you must verify that all of the core group members are running at a product code level (VRM) that is equal to or greater than the VRM in which the desired protocol version was added to the product. For example:bprac

Supported core group protocol version IDs

The following tables summarize for each protocol category the minimum level of the product that core group members must be running at before they can be associated with a specific protocol version. These tables also describe the new capabilities that were added in each protocol version.

Use these tables to determine which protocol versions you can use with a particular core group, and then use either the IBM_CS_WIRE_FORMAT_VERSION or the IBM_CS_HAM_PROTOCOL_VERSION core group custom property to configure all of the members of that core group to run using the newest of those protocol versions that is supported by the level of the product on which you are running. The high availability manager automatically detects the configuration changes and starts to use the new core group protocol version with these core group members.

Table 1. Supported wire format protocol version IDs. The protocol version ID indicates the first version, release, and modification level in which that version is included. The following table lists the supported wire format protocol version IDs.
Version ID Required Minimum product Level Description
6.0.0 Any This protocol version is the original or base version. All versions of the high availability manager can use this protocol. If you do not specify a particular wire format protocol version, the high availability manager uses this version.
6.0.2.9 6.0.2.9 for Version 6.0.2, and 6.1.0 for Version 6.1. This protocol version was included in the 6.0.2.9 service pack, and in Version 6.1 of the product to facilitate core group bridge scalability. This version is recommended for large topologies that contain multiple core groups and core group bridges as part of their configuration.
6.1.0 6.1.0 for Version 6.1 This protocol version was included in Versions 6.1.0 of the product. This version adds core group scalability improvements, and more support for large topologies.
Table 2. Supported high availability manager protocol version IDs . The protocol version ID indicates the first the version, release, and modification level in which that version is included. The following table lists the supported high availability manager protocol version IDs.
Version ID Required Minimum product Level Description
6.0.2.31 6.0.2.31 for Version 6.0.2, 6.1.0.19 for Version 6.1 This protocol version is the original or base version of the high availability manager protocol, and is available in the 6.0.2.31, and higher, and 6.1.0.19, and higher versions of the product to facilitate core group bridge scalability. This protocol version is recommended for topologies that contain multiple core groups and core group bridges as part of their configuration.



Related concepts
Core group Discovery Protocol
Core group Failure Detection Protocol
Core group View Synchrony Protocol
Core groups (high availability domains)
Related tasks
Selecting the version of a core group protocol
Concept topic    

Terms of Use | Feedback

Last updated: Feb 25, 2009 9:32:38 AM CST
http://publib.boulder.ibm.com/infocenter/wasinfo/v6r1/index.jsp?topic=/com.ibm.websphere.nd.multiplatform.doc/info/ae/ae/crun_ha_protocol_ver.html