Servers List

To see more details about a persistence service, go to the top bar of the FTL GUI, click Clients then Clusters and, finally, click the row in the Persistence Clusters status table to see the Servers List.

Columns

GUI Item Description
Server Status Status of the persistence service process (as a string).
State If the service is running, this string indicates its state or role with respect to the quorum and disaster recovery. For explanations of these state values, see Persistence Service States later in this topic.
Name The name of the persistence service.

This name originates in the realm definition, and is required in the FTL server configuration file.

ID Client ID of the persistence service (as a client of the realm service).

Administrators could use this client ID to retrieve client monitoring metrics about the persistence service using the FTL server web API.

Set The name of the disaster recovery set to which this service belongs.
Host The host name string of the persistence service’s host computer (if available).
History This field indicates the data state of a persistence service.

The first numeric value is the service’s quorum number.

The second numeric value is the data state of the service. Higher values indicate more recent data.

A history value of 0,0 indicates an empty service.

Consistency This field indicates whether a service is up to date.
  • Up to date: This service is up to date with respect to the other services in the cluster.
  • A history value pair indicates the quorum number and data state when this service was most recently up to date with respect to the other services in the cluster.

When the cluster does not contain enough up to date services to form a quorum, administrators may force a quorum, using the most up to date service (see Before Forcing a Quorum).

Disk When the service loads a state file from disk, or saves its state to disk, this JSON attribute reflects the state of the load or save operation.
Disk Persistence None
Disk persistence disabled.
sync
The client returns from a send-message call after the message has been written to a majority of disks. This mode generally provides consistent data and robustness, but at the cost of increased latency and lower throughput. If the cluster restarts, no data is lost; performance is subject to disk performance.
async
The client may return from a send-message call before the message has been written to disk by majority of the FTL servers. This mode generally provides less latency and more throughput, but messages could be lost if a majority of servers restart shortly after the API call.
Disk Swap

Message swapping is enabled (true) or disabled (false).

When message swapping is enabled (with or without disk persistence), message data is freed from persistence service process memory according to configured limits. However, the system may choose to use idle memory as a cache for the message data on disk. Tools used to report the memory usage of the persistence service process may include this cache memory, which is normal behavior. If the system experiences memory pressure, it reclaims the cache memory for use in other processes.

Commands

Icons at the right of each cluster row trigger commands:

  • Change logging level of a persistence service.

    For the meaning of log level values, see "Log Level Reference" in TIBCO FTL Development.

  • Save the persistence service's data state to a file.

    This command icon is visible only when the command is available. For background information, see Saving and Loading Persistence State and its subtasks.

Persistence Service States

Role Description
Leader Functioning properly. This service is the leader.
Replica Functioning properly. This service is a replica.
Offline This persistence service is configured, but not running.

The persistence service or the FTL server that provides it might have exited, or a network segmentation prevents connection.

Forming Quorum This process is running, but the service is not yet functioning as part of a valid quorum.

If this state continues, ensure that all persistence services are running, and check network connectivity.

Timed Out The FTL server no longer detects this persistence service.
Manual Intervention Potential danger of data loss.

This service is the provisional leader, but it cannot safely form a quorum even though other services are reachable.

To proceed, see Before Forcing a Quorum.

Needs Restart Warning.

The administrator is testing a deployment with configuration changes that would require this persistence service to restart. Meanwhile, this service is still functioning properly using the old realm definition.

Out of Sync Error.

This persistence service is still functioning, but its realm definition is not the most recent deployment.

(This situation can occur when a network partition separates persistence services. When the network reconnects the FTL server GUI could alert you to stop the persistence service so that the FTL server automatically restarts it with the newer deployment.)

Exception Recoverable error.

A non-fatal error occurred. After briefly registering this status, the persistence service will attempt to form a quorum.

Check log files to determine the reason for the error.

DR Leader Functioning properly. This service is the disaster recovery leader in the standby set.

See Persistence Service Sets: Primary and Standby

DR Replica Functioning properly. This service is a disaster recovery replica in the standby set.

See Persistence Service Sets: Primary and Standby

Also see Catalog of Persistence Metrics