Basic Commands
Basic cluster management commands using the command line.
This guide covers the basic commands for managing various services in your PostgreSQL Cluster. It includes instructions for working with Patroni, PostgreSQL, PgBouncer, etcd, and other essential services.
For more detailed information about the cluster's architecture and components, please refer to the Architecture page.
Patroni
Official documentation
Service
To check the patroni service status:
sudo systemctl status patroni
Other actions: start, stop, restart, reload:
sudo systemctl [start|stop|restart|reload] patroni
When executing commands such as stop, start, or restart, Patroni also restarts the Postgres service, unless the Patroni cluster is in maintenance mode.
logs
View Patroni service logs (last 50 lines, with live updates):
sudo journalctl -u patroni -n 50 -f
If you use logging to a file (variable patroni_log_destination: logfile):
sudo tail -n 50 -f /var/log/patroni/patroni.log
CLI (patronictl)
For more information on all available options, run patronictl --help. Or refer to the patronictl documentation.
List the cluster members:
patronictl list
Auto update the screen every 2 seconds and print timestamp:
patronictl list -t -W
Show cluster configuration:
patronictl show-config <cluster-name>
Edit cluster configuration:
patronictl edit-config <cluster-name>
Reload cluster configuration:
patronictl reload <cluster-name>
Reload for specific cluster member:
patronictl reload <cluster-name> <server-name>
Restart cluster:
patronictl restart <cluster-name>
Restart specific cluster member:
patronictl restart <cluster-name> <server-name>
This command restarts the Postgres service but does not affect the Patroni service. If you need to restart the Patroni service, use the systemctl restart patroni command.
Switchover to a replica:
patronictl switchover <cluster-name>
This is typically done during scheduled maintenance, for example, when you need to perform maintenance on a server currently holding the "leader" role.
Failover to a replica:
patronictl failover <cluster-name>
Usually, this command is not necessary, as Patroni will handle failover automatically. However, it may be required in certain situations, such as when there is no suitable candidate for the leader role (e.q, when the lag specified in the maximum_lag_on_failover option is exceeded). Use this command with caution.
Show the history of failovers/switchovers:
patronictl history <cluster-name>
Disable auto failover (start maintenance mode):
patronictl pause <cluster-name> --wait
Maintenance mode is an option in Patroni that allows you to pause automatic failover.
This can be useful in scenarios such as performing a PostgreSQL upgrade or during DCS maintenance, where you may need to temporarily disable automatic cluster management. Without this mode, if the master becomes unavailable, Patroni will automatically promote a replica to the leader role.
Enable auto failover (stop maintenance mode):
patronictl resume <cluster-name> --wait
Reinitialize cluster member:
patronictl reinit <cluster-name> <server-name>
This command is used when you need to reinitialize the database data files on a replica, for example, if it has fallen significantly behind (high replocation lag). Patroni will delete the data directory and re-create it using the methods defined in the create_replica_methods option.
PostgreSQL
Official documentation
Service
In a clustered environment, there is no need to manage PostgreSQL services separately, as Patroni handles this automatically. See Patroni "Service" section.
logs
View the list of log files:
sudo ls -lth /var/log/postgresql/
View current Postgres log (last 50 lines, with live updates):
sudo tail -n 50 -f $(ls -t /var/log/postgresql/*.log | head -n 1)
CLI (psql)
For more information on all available options, run psql --help. Or refer to the psql documentation.
Connect to database "postgres":
psql -h 127.0.0.1 -p 5432 -U postgres -d postgres
Execute the SQL command:
psql -h 127.0.0.1 -p 5432 -U postgres -d postgres -c "select version()"
PgBouncer
Official documentation
Service
To check the pgbouncer service status:
sudo systemctl status pgbouncer
Other actions: start, stop, restart, reload:
sudo systemctl [start|stop|restart|reload] pgbouncer
logs
View pgbouncer service logs (last 50 lines, with live updates):
sudo tail -n 50 -f /var/log/pgbouncer/pgbouncer.log
CLI (psql)
For more information on all available options, run SHOW HELP. Or refer to the documentation.
Connect to pgbouncer database:
psql -h 127.0.0.1 -p 6432 -U postgres -d pgbouncer
Show databases:
SHOW DATABASES;
Show pools:
SHOW POOLS;
Show stats:
SHOW STATS;
HAProxy
Official documentation
Service
To check the haproxy service status:
sudo systemctl status haproxy
Other actions: start, stop, restart, reload:
sudo systemctl [start|stop|restart|reload] haproxy
logs
View haproxy service logs (last 50 lines, with live updates):
sudo journalctl -u haproxy -n 50 -f
sudo tail -n 50 -f /var/log/haproxy.log
Stats
To view HAProxy statistics, navigate to the following URL in your web browser:
http://<server-ip>:7000
confd
Used to automate HAProxy configuration file management.
Official documentation
Configuration files
/etc/confd/confd.toml
/etc/confd/conf.d/haproxy.toml
/etc/confd/templates/haproxy.tmpl
To update HAProxy configuration, edit the template located at: /etc/confd/templates/haproxy.tmpl
Then reload the confd service. Changes will be applied automatically during the next update.
Service
To check the confd service status:
sudo systemctl status confd
Other actions: start, stop, restart, reload:
sudo systemctl [start|stop|restart|reload] confd
logs
View confd service logs (last 50 lines, with live updates):
sudo journalctl -u confd -n 50 -f
Keepalived
Official documentation
Service
To check the keepalived service status:
sudo systemctl status keepalived
Other actions: start, stop, restart, reload:
sudo systemctl [start|stop|restart|reload] keepalived
logs
View keepalived service logs (last 50 lines, with live updates):
sudo journalctl -u keepalived -n 50 -f
CLI (ip)
Check the IP address list:
ip a
vip-manager
Official documentation
Service
To check the vip-manager service status:
sudo systemctl status vip-manager
Other actions: start, stop, restart, reload:
sudo systemctl [start|stop|restart|reload] vip-manager
When the vip-manager service is stopped, the VIP address is removed. It will be re-added when the service starts again, provided that the current server is the leader.
logs
View vip-manager service logs (last 50 lines, with live updates):
sudo journalctl -u vip-manager -n 50 -f
CLI (ip)
Check the IP address list:
ip a
etcd
Official documentation: Operations guide
Service
To check the etcd service status:
sudo systemctl status etcd
Other actions: start, stop, restart, reload:
sudo systemctl [start|stop|restart|reload] etcd
logs
View etcd service logs (last 50 lines, with live updates):
sudo journalctl -u etcd -n 50 -f
CLI (etcdctl)
For more information on all available options, run etcdctl --help
Lists all members in the cluster:
etcdctl member list
Check the cluster health:
etcdctl endpoint health --cluster
Prints out the status of cluster endpoints:
etcdctl endpoint status --cluster -w table
Consul
Service
To check the consul service status:
sudo systemctl status consul
Other actions: start, stop, restart, reload:
sudo systemctl [start|stop|restart|reload] consul
logs
View Consul service logs (last 50 lines, with live updates):
sudo journalctl -u consul -n 50 -f
CLI (consul)
For more information on all available options, run consul --help
List all members in the cluster:
consul members
Check the health of the cluster:
consul operator raft list-peers
Print out the status of all services in the cluster:
consul catalog services
List all nodes running a specific service (e.g., postgres-cluster):
consul catalog nodes -service=<cluster-name>
CLI (dig)
Resolve DNS for the master node of the cluster:
dig @127.0.0.1 -p 8600 +short master.<cluster-name>.service.consul SRV
Resolve DNS for the replica nodes of the cluster:
dig @127.0.0.1 -p 8600 +short replica.<cluster-name>.service.consul SRV