Persistence: Stores and Durables

Stores and durables provide for persisted messages and persistent interest.

For an intuitive introduction and terminology, see “Persistence: Stores and Durables” in TIBCO FTL Concepts.

  • Purposes of Persistence
    The persistence infrastructure of stores and durables can serve four purposes: delivery assurance, apportioning message streams, last-value availability, and key/value maps. These purposes correspond to four types of durables: standard durables, shared durables, last-value durables, and map durables.
  • Persistence Architecture
    Persistence is flexible. Administrators can tailor various aspects of persistence to meet the needs of applications.
  • Coordination for Persistence
    Administrators and application developers must coordinate the details of durables.
  • Stores for Delivery Assurance
    The following topics present the details of delivery assurance for architects and administrators.
  • Stores for Apportioning Message Streams
    This topic presents the details of apportioning message streams for architects and administrators.
  • Stores for Last-Value Availability
    This topic presents the details of last-value availability for architects and administrators.
  • Stores for Message Broker
    This topic presents information for architects and administrators about using stores to implement a message broker.
  • Default Durable
    You can use the default durable to add delivery assurance to any application, even applications that do not explicitly supply a durable name or a subscriber name in their create subscriber calls.
  • Durable Behavior
    Administrators configure the behavior of a durable and its interaction with durable subscribers.
  • Wide-Area Forwarding
    Client applications can publish and subscribe across network boundaries. To enable this behavior, administrators define forwarding zones and wide-area stores.
  • Arranging a Wide-Area Store
    To define a wide-area store, use the stores grid of the FTL server GUI.
  • Wide-Area Forwarding: Runtime Conditions
    Wide-area forwarding places configuration prerequisites on the runtime topology of FTL servers, and clusters of persistence services.
  • Persistence Services and Clusters
    If a cluster of FTL servers provide persistence services, those persistence services cooperate in a persistence cluster. A persistence cluster maintains a set of persistence stores which store messages in durables.
  • Configuring Persistence
    The realm definition schematic in the following diagram indicates five configuration tasks for persistence stores and durables. (Numbers in the diagram correspond to the task steps that follow.)
  • Memory Reserve for Persistence Services
    Memory reserve is pre-allocated memory with which a persistence service could continue limited operations after exhausting available memory.
  • Starting a Persistence Service
    Administrators have full responsibility for persistence service configuration and operation.
  • Stopping or Restarting a Persistence Service
    To permanently stop an individual persistence service, terminate the FTL server process that manages it. To stop and restart an individual persistence service, use the shutdown command.
  • Persistence Monitoring and Management
    The following topics present tools in the FTL server GUI to help monitor and manage persistence.
  • Before Forcing a Quorum
    When a persistence cluster contains enough reachable persistence services, but not enough non-empty services to form a quorum, an administrator may force a quorum. Read this topic first.
  • Saving and Loading Persistence State
    The persistence GUI includes a facility to save the content of a persistence service’s stores to a file. When a new persistence service starts, it can load its initial state from that file. You can use this facility for nightly backups, during upgrade migration, or to migrate a persistence cluster to a new physical location.
  • Clock Synchronization, Affiliated FTL Servers, and Persistence
    Persistence services depend on clock synchronization among their host computers.