Cluster start-stop sequence

This refers to the clustered version of the service start-stop.

System startup sequence

1. Start adbmgr.

mgr_ctl start -D [the data directory of adbmgr] -l [Path to log file]

For example:

mgr_ctl start -D /home/uatadb/mgr -l /home/uatadb/mgr/logfile

2. Connecting to adbmgr using psql.

psql -d postgres -p [the port number of adbmgr]

For example:

psql -d postgres -p 31000

3. Start agent.

• Start all agent:

start agent all password ‘[agent login password for the installation user]’;

For example:

start agent all password '123456';

The above command is executed successfully provided that the passwords of the installation users on all hosts are the same; if the passwords of the agent installation users are different, you need to start the agents one by one. The command is as follows.

• Start an agent.

start agent [agent name] password ‘[agent login password for the installation user]’;

For example:

 start agent host1 password '123456';

4. Start the nodes in the cluster.

• Start all nodes.

start all;

• Start the specified node.

start {gtmcoord|coordinator|datanode} {master|slave} [node name]

For example:

start gtmcoord master start；

System stop sequence

1. Connecting to adbmgr using psql.

psql -d postgres -p [the port name of adbmgr]

For example:

psql -d postgres -p 31000

2. Stop all nodes in the cluster.

• Stop all nodes.

stop all;

• Stop specified nodes.

stop{gtmcoord|coordinator|datanode} {master|slave} [node name]

For example:

stop gtmcoord master start；

3. Stop the agent.

• Stop all agents.

stop agent all;

• Stop individual agents.

stop agent [agent name];

4. Stop adbmgr.

mgr_ctl stop-D [the data directory of adbmgr]

For example:

mgr_ctl stop -D /home/uatadb/mgr