Configuration: Cluster

The cluster module permits the clustering of Transformers allowing data to be replicated between Transformers.

Data can be replicated either by programmatically or by configuration.

Loading the Cluster module

To load the cluster module add the following to transformer.conf

add-module  cluster

Options

The following options should be placed in cluster.conf and configure the way that replication takes place.

connect-sync-ignore

Type: String Array
Default: None

Symbol patterns ignore when connecting

connect-sync

Type: String Array
Default: None

Symbol patterns to automatically sync on connect

always-auto-distribute-ignore

Type: String Array
Default: None

Symbol patterns ignore when distributing regardless of cluster state

always-auto-distribute

Type: String Array
Default: None

Symbol patterns to automatically distribute regardless of cluster state

auto-distribute-ignore

Type: String Array
Default: None

Symbol patterns ignore when distributing

auto-distribute

Type: String Array
Default: None

Symbol patterns to automatically distribute

cluster-index

Type: Integer
Default: -1
Minimum: 0

Index of this node

Configuration group add-cluster-node

Add a cluster node

cluster-addr

Type: String
Default: 127.0.0.1

Address of cluster node

cluster-port

Type: Integer
Default: -1

Port of cluster node

default-primary

Type: Boolean
Default: FALSE

Default primary

force-primary

Type: Boolean
Default: FALSE

Force this instance to be primary

primary-config-file

Type: String
Default: %r/etc/cluster-primary.conf

Config file for the forced status

cluster-packet-logfile

Type: String
Default: None

Log cluster packets into this file

primary-method

Type: String
Default: order

Method of determining primary mode operation

stable-time

Type: Integer
Default: 5

Time for system to be up before becoming primary

logfile

Type: String
Default: cluster.log

Logfile

log-level

Type: String
Default: INFO

Logging level of the pipeline module.

check-period

Type: Floating Point Value
Default: 1.000000

Interval to check primary status at (monitoring only)

connect-sync-always

Type: Boolean
Default: FALSE

Always sync data regardless of primary status

cluster-connect-timeout

Type: Floating Point Value
Default: 10.000000

Timeout for connections

cluster-nosave-on-secondary

Type: Boolean
Default: FALSE

Don't save replicated symbols on secondary

heartbeat-time

Type: Integer
Default: 15
Minimum: 0

Time between sending heartbeats

heartbeat-slack-time

Type: Integer
Default: 5
Minimum: 0

Max time between reading heartbeats