Implementation of Locking by Using a Store

Earlier locking was implemented as part of the cluster provider: FTL, Ignite, or Legacy ActiveSpaces. Now you can implement locking by using a store. Store-based locking provides a single consistent way of locking irrespective of cluster and store provider combination. Store-based locking can be used in unclustered environments.

Locking Scenarios

To determine which locking implementation is used in the respective scenarios, refer to the following table:

Clustered

Cluster Provider	Cache	Store	Locking
Apache Ignite	Apache Ignite	Yes	Apache Ignite or Store
TIBCO FTL	Apache Ignite	Yes	Apache Ignite or Store
TIBCO FTL	NA	Yes	Store

Unclustered

In the unclustered mode, if the Store option is selected then the store-based locking is used.

Configuration Properties

On the Cluster tab of the CDD file, to enable store-based locking, configure the following properties.

Property Name	Property Value	Description
be.engine.store.locking.enabled	True or False	Enables or disables the store locking feature.
be.engine.api.spec.enable	True or False	Enables or disables the MemberHealthStatus API. This API is used to determine whether a member is dead or alive.
be.engine.http.ping.port	Port range	Specifies the port number range on which the MemberHealthStatus API would be available. Specify the lower and upper port range, separated by hyphen ( - ). Example: 8180-8190

Apache Ignite Locking Properties

Due to network issues, there may be a failure in locking because of disconnection. To ensure store based locking for Apache Ignite, on the Cluster tab of the CDD file, configure the following properties.

Property Name	Description
be.ignite.lock.disconnect.retry.interval	Specifies the interval duration to retry locking when disrupted due to disconnection. The default value is 5 seconds.
be.ignite.lock.disconnect.retry.attempts	Specifies the number of attempts allowed to retry locking when disrupted due to disconnection. The default value is 3.