Replicate instances and services on Linux
Replicate supports running multiple instances concurrently on the same Linux server with a single installation. This may be useful, for instance, if you need to run several tasks or groups of tasks, but with different Replicate server settings for each group or task.
Instances can be installed as Linux services, which means that the instance will be stopped in an orderly fashion when a server is shut down and restarted when a server is rebooted. Installing Replicate creates an initial instance, named areplicate, which is installed as a service.
Additional instances can be created after Replicate is installed. An additional instance will not be installed as a service if installed by a non-root user. A new instance will not be run when it is created, allowing for configuration first.
Replicate instances get their environment from several files:
| File | Description |
|---|---|
| arep_login.sh | Sets shell variables, e.g. LD_LIBRARY_PATH for all instances. This file should not be modified manually. |
| site_arep_login.sh | Created functionally null; can be modified manually to contain shell settings for all instances. |
| instancename_arep_login.sh | Created functionally null in the instance's data directory, for an instance-specific shell configuration. |
Each instance has a unique name, uses its own ports, and has its own data directory. This means that when running commands such as repctl, you need to specify the instance-specific data directory in the command line so that the correct instance is affected.
-
When starting and stopping an instance, it is strongly suggested to use the instance script rather than run repctl directly.
-
All instances are run as the same user. Instances are listed in a file named services_list.txt in the Replicate root directory. This file should not be modified manually.
-
When upgrading Replicate, all instances will be reinstalled and started, even if they were not running prior to the upgrade.