Beginning with version 1.6.6 of dCache the replica manager can be started as follows:

The replica manager will use the same PostgreSQL database and database user srmdcache as the SRM. The standard configuration assumes that the database server is installed on the same machine as the replica manager — usually the admin node of the dCache instance. To create and configure the database replicas used by the replica manager in the database server do:

[root] # su postgres
[user] $ createdb -U srmdcache replicas
[user] $ psql -U srmdcache -d replicas -f /opt/d-cache/etc/psql_install_replicas.sql
[user] $ exit

The start-up script bin/dcache-core already contains the correct lines to start and stop the domain containing the replica manager as comments. Just remove the two hash (“#”) signs and restart the dCache instance. The replica manager may also be started separately by

[root] # /opt/d-cache/jobs/replica -logfile=/opt/d-cache/log/replica.log start

and stopped by

[root] # /opt/d-cache/jobs/replica stop

In the default configuration, all pools of the dCache instance will be managed. The replica manager will keep the number of replicas between 2 and 3 (including). At each restart of the replica manager the pool configuration in the database will be recreated.