Clustering configuration

The following configuration items define how Liberator participates in a cluster.

Want to know more about clustering? See here.

Note: "this Liberator" means the Liberator for which you are defining the configuration. "this node" means the cluster node that this Liberator represents.

Contents:

add-cluster-node

add-cluster-node identifies a Liberator in the cluster.

Tip: For more about how to use add-cluster-node, see Clustering Liberators without the Caplin Deployment Framework in How can I.. Set up a cluster of Liberators.

Syntax:

add-cluster-node
   cluster-addr [string]
   cluster-port [integer]
end-cluster-node
Option Type Default Description
cluster-addr string 127.0.0.1

The network interface address or hostname of the cluster node.

In the Caplin Platform Deployment Framework, the configuration macros LIBERATOR${THIS_LEG}_HOST and LIBERATOR${OTHER_LEG}_HOST specify the cluster-addr values. Clustering is used in the Deployment Framework to support failover capability. You normally use the  ./dfw hosts command to set up the hostnames in the macros. See Configuring server hostnames in How can I... Set up server failover capability, and Change server hostnames in How can I ... Change server-specific configuration.

Also see Deployment Framework Configuration macros and items.

cluster-port integer 2333

The network port number of the cluster node. This port is used for cluster-related communication between the Liberator and the other Liberators in the cluster.

In the Caplin Platform Deployment Framework, the configuration macros LIBERATOR${THIS_LEG}_CLUSTER_PORT and LIBERATOR${OTHER_LEG}_CLUSTER_PORT specify the cluster-port values. See Clustering Liberators in the Caplin Deployment Framework in How can I ... Set up a cluster of Liberators.

Also see Deployment Framework Configuration macros and items.

cluster-cache-request-objects

cluster-cache-request-objects specifies, when set to TRUE, that this Liberator is to automatically request (subscribe to) the objects that the other Liberators in the cluster request.

If you set this option to TRUE for one Liberator, make it TRUE for all the other Liberators in the cluster. Then all the Liberators will request (and cache) an object when one of them requests it.

Syntax: cluster-cache-request-objects TRUE

Type: boolean

Default value: FALSE

cluster-cache-source-routing

cluster-cache-source-routing specifies, when set to TRUE, that this Liberator is to request objects from the same DataSource application instance (Transformer or Integration Adapter) as other Liberators in the cluster that request those objects. Setting this item enables Intelligent Source Routing.

If you set this option to TRUE for one Liberator, make it TRUE for all the other Liberators in the cluster. Then all the Liberators will share information about which DataSource application an object was requested from.

Enabling cluster cache source routing can have two advantages. Firstly it minimises the load on the Transformers/Integration Adapters because they aren't all serving up the same objects. Secondly it minimises the bandwidth used on the network connecting the Transflormers/Integration Adapters to the Liberators, as each update is only sent by a single Transformer/Integration Adapter - this can be a significant advantage if the Transformers/Integration Adapters are connected to the Liberators over a WAN.

Syntax: cluster-cache-source-routing TRUE

Type: boolean

Default value: FALSE

There are situations where Intelligent Source Routing may not be enforced:

  • If two Liberators both make an initial request for an object at around the same time, there may not be enough time for the Liberator that first received the request to tell the second Liberator which DataSource application instance to request the object from; the second Liberator may already have requested the object from a different DataSource application instance. 
  • Intelligent Source Routing may not always be maintained in failover situations. For example, if multiple Liberators are receiving updates from an Integration Adapter that subsequently goes down, they will failover to another Integration Adapter. But they may each failover to a different Adapter and after this they will continue to receive updates for existing subscriptions from their respective new Adapters. However, after the failover, any new subscriptions will be subject to Intelligent Source Routing.

cluster-heartbeat-slack-time

cluster-heartbeat-slack-time When this Liberator doesn't receive an expected cluster heartbeat (see cluster-heartbeat-time) from another Liberator in the cluster, it waits cluster-heartbeat-slack-time seconds before disconnecting from the cluster and trying to reconnect to it.

Unlike cluster-heartbeat-time, the value of cluster-heartbeat-slack-time is not compared by the Liberators in the cluster.

Syntax: cluster-heartbeat-slack-time <slack-time-in-seconds>

Type: integer

Default value: 5 seconds

cluster-heartbeat-time

cluster-heartbeat-time specifies the time in seconds between each cluster heartbeat. The Liberators forming a cluster exchange heartbeat messages at regular intervals, allowing each of them to check that all the others are still present. If the cluster heartbeat times out, this Liberator waits cluster-heartbeat-slack-time and then disconnects from the cluster and tries to reconnect to it.

When all the Liberators in the cluster have initially connected, they compare their cluster-heartbeat-time values and use the lowest.

In the Caplin Platform Deployment Framework, you use a configuration macro LIBERATOR_CLUSTER_HEARTBEAT_TIME to specify Liberator's cluster-heartbeat-time. See Clustering Liberators in the Caplin Deployment Framework in How can I ... Set up a cluster of Liberators, and Deployment Framework Configuration macros and items.

Syntax: cluster-heartbeat-time <heartbeat-time-in-seconds>

Type: integer

Default value: 15 seconds

The minimum allowed value is 1 second.

cluster-index

cluster-index specifies the index number of this cluster node. The index number uniquely identifies this Liberator within the cluster. It states which of the add-cluster-node items this node represents. Index numbers start at 0 and correspond to the order of the add-cluster-node items in the configuration file.

In the Caplin Platform Deployment Framework, you use a configuration macro LIBERATOR_CLUSTER_NODE_INDEX to specify Liberator's cluster-index. See Clustering Liberators in the Caplin Deployment Framework in How can I ... Set up a cluster of Liberators, and Deployment Framework Configuration macros and items.

Syntax: cluster-index <index-number>

Type: integer

Default value: 0

cluster-packet-logfile

cluster-packet-logfile specifies the name of the log file in which message packets passed between this Liberator and other nodes in the Liberator cluster are recorded.

The filename can contain the parameter %a, which is replaced at run-time by this Liberator's DataSource application-name. The log file is located in Liberator's var directory (the <Framework-root>/servers/Liberator/var directory in the Deployment Framework).

The log file's format is the same binary format used for the standard DataSource packet log (see datasrc-pkt-log in DataSource peers configuration part 2), so you need to use the logcat utility to read it.

Warning: only configure cluster-packet-logfile when you need to troubleshoot your Liberator cluster. Cluster packet logs can quite quickly become very large, so you should only enable logging of this information for debugging purposes.

Syntax: cluster-packet-logfile <log-file-name>

Type: string

Default value: (none, but this configuration item is optional - if it's not present then cluster packets aren't logged)


See also: