Starting and Stopping Tungsten Components

To stop all of the services associated with a dataservice node, use the stopall script:

shell> stopall 
Stopping Tungsten Connector...
Stopped Tungsten Connector.
Stopping Tungsten Replicator Service...
Stopped Tungsten Replicator Service.
Stopping Tungsten Manager Service...
Stopped Tungsten Manager Service.

To start all services, use the startall script:

shell> startall
Starting Tungsten Manager Service...
Starting Tungsten Replicator Service...
Starting Tungsten Connector...

Stopping/Starting Tungsten Replicator

warning

Stopping, or taking offline, a running replicator temporarily stops and restarts replication. Either set MAINTENANCE mode within cctrl (see "operations-maintenance") or shun the datasource before stopping the replicator if you need to ensure it does not restart.

warning

Before stopping a replicator it is good practice to take it offline cleanly first. Doing so will ensure a clean stop of any replication in flight and ensure replication positions are correctly recorded.

If Parallel Apply is in use, this is essential, see "Parallel Replication and Offline Operation" for more information

To shutdown a running Tungsten Replicator you must switch off the replicator:

shell> trepctl -all-services offline
shell> replicator stop
Stopping Tungsten Replicator Service...
Stopped Tungsten Replicator Service.

To start the Tungsten Replicator service if it is not already running:

shell> replicator start
Starting Tungsten Replicator Service...

To restart the Tungsten Replicator service if it is running:

shell> trepctl -all-services offline
shell> replicator restart
Stopping Tungsten Replicator Service...
Stopped Tungsten Replicator Service.
Starting Tungsten Replicator Service...

To start the Tungsten Replicator service if it is not already running, but ensure replication is offline:

shell> replicator start offline
Starting Tungsten Replicator Service...

Stopping/Starting Tungsten Connectors

warning

Restarting the connector service will interrupt the communication of any running application or client connecting through the connector to MySQL.

To shutdown a running Tungsten Connector:

shell> connector stop
Stopping Tungsten Connector Service...
Stopped Tungsten Connector Service.

To start the Tungsten Connector service if it is not already running:

shell>  connector start
Starting Tungsten Connector Service...
Waiting for Tungsten Connector Service.....
running: PID:12338

To restart the Tungsten Connector service if it is running:

shell> connector restart
Stopping Tungsten Replicator Service...
Stopped Tungsten Replicator Service.
Starting Tungsten Replicator Service...

If the cluster was configured with auto-enable=false then you will need to put each node online individually.

Stopping/Starting Tungsten Managers

The manager service is designed to monitor the status and operation of the each of the datasources within the dataservice. In the event that the manager has become confused with the current configuration, for example due to a network or node failure, the managers can be restarted. This forces the managers to update their current status and topology information.

Before restarting managers, the dataservice should be placed in MAINTENANCE policy mode. In MAINTENANCE mode, the connectors will continue to service requests and the manager restart will not be treated as a failure.

To restart the managers across an entire dataservice, each manager will need to be restarted. The dataservice must be placed in MAINTENANCE policy mode first, then:

1. To set the MAINTENANCE policy mode:

   [LOGICAL:EXPERT] /alpha > set policy maintenance

2. On each datasource in the dataservice:

   1. Stop the service:

      shell> manager stop

   2. Then start the manager service:

      shell> manager start

   3. Or:

      shell> manager restart

3. Once all the managers have been restarted, set the policy mode back to AUTOMATIC:

   [LOGICAL:EXPERT] /alpha > set policy automatic 
   policy mode is now AUTOMATIC

Restarting the Multi-Site Active/Active Cross-Site Replicator Service

warning

Restarting a running replicator temporarily stops and restarts replication. When using Multi-Site/Active-Active, restarting the additional replicator will stop replication between sites.

These instructions assume you have installed the additional replicator with the executable-prefix option.

If not, you should go to /opt/replicator/tungsten/tungsten-replicator/bin and run the replicator and trepctl commands directly.

warning

Before stopping a replicator it is good practice to take it offline cleanly first. Doing so will ensure a clean stop of any replication in flight and ensure replication positions are correctly recorded.

If parallel Apply is in use, this is essential, see "Parallel Replication and Offline Operation" for more information.

warning

The Tungsten Replicator services that connect clusters in a Multi-Site/Active-Active topology are not managed by the individual cluster services, and therefore the cluster will not automatically restart the cross-site replicator services.

To shutdown a running Tungsten Replicator you must switch off the replicator:

shell> mm_trepctl -all-services offline
shell> mm_replicator stop
Stopping Tungsten Replicator Service...
Stopped Tungsten Replicator Service.

To start the Tungsten Replicator service if it is not already running:

shell> mm_replicator start
Starting Tungsten Replicator Service...

To restart the Tungsten Replicator service if it is running:

shell> mm_trepctl -all-services offline
shell> mm_replicator restart
Stopping Tungsten Replicator Service...
Stopped Tungsten Replicator Service.
Starting Tungsten Replicator Service...

To start the Tungsten Replicator service if it is not already running, but ensure replication is offline:

shell> mm_replicator start offline
Starting Tungsten Replicator Service...