patronictl

Command reference for patronictl configuration, syntax, and subcommands.

Source: https://patroni.readthedocs.io/en/latest/patronictl.html

The patronictl command-line interface communicates with the Patroni REST API and the DCS. Use it for cluster operations, interactively or in scripts.

Configuration

patronictl uses 3 sections of the configuration:

ctl: REST API authentication and server identity verification. See ctl settings.
restapi: fallback settings when ctl does not supply REST API authentication or server verification. patronictl uses restapi.authentication if ctl.authentication is absent, and restapi.cafile if ctl.cacert is absent. See REST API settings.
DCS, such as etcd: DCS connection and authentication settings.

Supply these options through environment variables or a configuration file. See the corresponding sections in Environment Configuration Settings and YAML Configuration Settings.

Patronictl reads environment variables and uses their values directly.

By default, patronictl looks for patronictl.yaml in one of these system-specific directories. You can also specify a different file.

Mac OS X: ~/Library/Application Support/patroni
Mac OS X (POSIX): ~/.patroni
Unix: ~/.config/patroni
Unix (POSIX): ~/.patroni
Windows (roaming): C:\Users\<user>\AppData\Roaming\patroni
Windows (not roaming): C:\Users\<user>\AppData\Local\patroni

You can override that behavior either by:

Setting the environment variable PATRONICTL_CONFIG_FILE with the path to a custom configuration file;
Using the -c / --config-file command-line argument of patronictl with the path to a custom configuration file.

Note

On the patroni host, patronictl can use the daemon’s configuration file if it contains all required sections.

Usage

This section describes the operations available through patronictl.

The patronictl command accepts these global arguments:

-c / --config-file
As explained before, used to provide a path to a configuration file for patronictl.

-d / --dcs-url / --dcs
Provide a connection string to the DCS used by Patroni.

This argument sets or overrides the DCS and namespace in patronictl configuration.

Use DCS://HOST:PORT/NAMESPACE. For example, etcd3://localhost:2379/service connects to etcd v3 on localhost and stores cluster data under the service namespace. Missing components use configuration values or defaults.

-k / --insecure
Flag to bypass validation of REST API server SSL certificate.

This is the synopsis for running a command from the patronictl:

patronictl [ { -c | --config-file } CONFIG_FILE ]
  [ { -d | --dcs-url | --dcs } DCS_URL ] 
  [ { -k | --insecure } ]
  SUBCOMMAND

Note

Synopsis notation:

Square brackets enclose optional arguments.
Curly brackets enclose alternatives; select one.
[, ... ] permits repetition.
Uppercase names require a value.

All patronictl subcommands use this notation. Replace SUBCOMMAND in the general synopsis with the relevant subcommand synopsis.

The following sections describe each patronictl command. Examples use postgres0.yml, postgres1.yml, and postgres2.yml from the Patroni GitHub repository.

patronictl demote-cluster

Synopsis

demote-cluster
  [ CLUSTER_NAME ]
  [ --host HOST ]
  [ --port PORT ]
  [ --restore-command RESTORE_COMMAND ]
  [ --primary-slot-name PRIMARY_SLOT_NAME ]
  [ --force ]

Description

patronictl demote-cluster converts a regular Patroni cluster into a standby cluster.

The command builds standby_cluster from the supplied remote primary connection options and adds it to dynamic configuration. It waits until the leader becomes a standby leader. Before the change, it prints the topology and requests confirmation unless --force is set.

At least one of --host, --port or --restore-command must be specified.

Parameters

CLUSTER_NAME: Name of the Patroni cluster.

If omitted, patronictl uses scope if configured.

--host: Address of the remote node.

--port: Port of the remote node.

--restore-command: Command to restore WAL records from the remote primary.

--primary-slot-name: Name of the replication slot on the remote node to use for replication.

--force: Flag to skip confirmation prompts when demoting the cluster.

Useful for scripts.

Examples

Demote the cluster to a standby cluster that follows a remote primary endpoint:

$ patronictl -c postgres0.yml demote-cluster batman --host 192.0.2.10 --port 5432 --primary-slot-name batman --force

patronictl dsn

Synopsis

dsn
  [ CLUSTER_NAME ]
  [ { { -r | --role } { leader | primary | standby-leader | replica | standby | any } | { -m | --member } MEMBER_NAME } ]
  [ --group CITUS_GROUP ]

Description

patronictl dsn gets the connection string for one member of the Patroni cluster.

If several members match, the command selects one and gives priority to the primary.

Parameters

CLUSTER_NAME: Name of the Patroni cluster.

If omitted, patronictl uses scope if configured.

-r / --role
Choose a member that has the given role.

Role can be one of:

leader: the leader of either a regular Patroni cluster or a standby Patroni cluster; or
primary: the leader of a regular Patroni cluster; or
standby-leader: the leader of a standby Patroni cluster; or
replica: a replica of a Patroni cluster; or
standby: same as replica; or
any: any role. Same as omitting this parameter; or

-m / --member
Choose a member of the cluster with the given name.

MEMBER_NAME is the name of the member.

--group
Choose a member that is part of the given Citus group.

CITUS_GROUP is the ID of the Citus group.

Examples

Get DSN of the primary node:

$ patronictl -c postgres0.yml dsn batman -r primary
host=127.0.0.1 port=5432

Get DSN of the node named postgresql1:

$ patronictl -c postgres0.yml dsn batman --member postgresql1
host=127.0.0.1 port=5433

patronictl edit-config

Synopsis

edit-config
  [ CLUSTER_NAME ]
  [ --group CITUS_GROUP ]
  [ { -q | --quiet } ]
  [ { -s | --set } CONFIG="VALUE" [, ... ] ]
  [ { -p | --pg } PG_CONFIG="PG_VALUE" [, ... ] ]
  [ { --apply | --replace } CONFIG_FILE ]
  [ --force ]

Description

patronictl edit-config changes cluster dynamic configuration and writes it to the DCS.

Note

In a TTY, the command displays the configuration diff through a pager. It tries less or more by default. Set PAGER to use another pager.

Parameters

CLUSTER_NAME: Name of the Patroni cluster.

If omitted, patronictl uses scope if configured.

--group
Change dynamic configuration of the given Citus group.

If omitted, patronictl uses citus.group if configured.

CITUS_GROUP is the ID of the Citus group.

-q / --quiet
Flag to skip showing the configuration diff.

-s / --set
Set a given dynamic configuration option with a given value.

CONFIG is the name of the dynamic configuration path in the YAML tree, with levels joined by . .

VALUE is the value for CONFIG. If it is null, then CONFIG will be removed from the dynamic configuration.

-p / --pg
Set a given dynamic Postgres configuration option with the given value.

This is shorthand for --s / --set with postgresql.parameters. before CONFIG.

PG_CONFIG is the name of the Postgres configuration to be set.

PG_VALUE is the value for PG_CONFIG. If it is null, then PG_CONFIG will be removed from the dynamic configuration.

--apply
Apply dynamic configuration from the given file.

It is similar to specifying multiple -s / --set options, one for each configuration from CONFIG_FILE.

CONFIG_FILE is the YAML file containing changes to apply. Use - to read from stdin.

--replace
Replace the dynamic configuration in the DCS with the dynamic configuration specified in the given file.

CONFIG_FILE is the YAML file containing replacement configuration. Use - to read from stdin.

--force
Flag to skip confirmation prompts when changing the dynamic configuration.

Useful for scripts.

Examples

Change max_connections Postgres GUC:

patronictl -c postgres0.yml edit-config batman --pg max_connections="150" --force
---
+++
@@ -1,6 +1,8 @@
loop_wait: 10
maximum_lag_on_failover: 1048576
postgresql:
+  parameters:
+    max_connections: 150
  pg_hba:
  - host replication replicator 127.0.0.1/32 md5
  - host all all 0.0.0.0/0 md5

Configuration changed

Change loop_wait and ttl settings:

patronictl -c postgres0.yml edit-config batman --set loop_wait="15" --set ttl="45" --force
---
+++
@@ -1,4 +1,4 @@
-loop_wait: 10
+loop_wait: 15
maximum_lag_on_failover: 1048576
postgresql:
  pg_hba:
@@ -6,4 +6,4 @@
  - host all all 0.0.0.0/0 md5
  use_pg_rewind: true
retry_timeout: 10
-ttl: 30
+ttl: 45

Configuration changed

Remove maximum_lag_on_failover setting from dynamic configuration:

patronictl -c postgres0.yml edit-config batman --set maximum_lag_on_failover="null" --force
---
+++
@@ -1,5 +1,4 @@
loop_wait: 10
-maximum_lag_on_failover: 1048576
postgresql:
  pg_hba:
  - host replication replicator 127.0.0.1/32 md5

Configuration changed

patronictl failover

Synopsis

failover
  [ CLUSTER_NAME ]
  [ --group CITUS_GROUP ]
  --candidate CANDIDATE_NAME
  [ --force ]

Description

patronictl failover performs a manual failover in the cluster.

Use this command when the cluster is unhealthy, for example:

There is no leader; or
There is no synchronous standby available in a synchronous cluster.

With synchronous mode enabled, this command can also fail over to an asynchronous node.

Note

patronictl failover also works on a healthy cluster. We recommend patronictl switchover for healthy clusters.

Warning

Failover can lose data if the promoted replica is behind the primary.

Parameters

CLUSTER_NAME: Name of the Patroni cluster.

If omitted, patronictl uses scope if configured.

--group
Perform a failover in the given Citus group.

CITUS_GROUP is the ID of the Citus group.

--candidate
The node to be promoted on failover.

CANDIDATE_NAME is the name of the node to be promoted.

--force
Flag to skip confirmation prompts when performing the failover.

Useful for scripts.

Examples

Fail over to node postgresql2:

$ patronictl -c postgres0.yml failover batman --candidate postgresql2 --force
Current cluster topology
+ Cluster: batman (7277694203142172922) -+-----------+----+-------------+-----+------------+-----+
| Member      | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
| postgresql0 | 127.0.0.1:5432 | Leader  | running   |  3 |             |     |            |     |
| postgresql1 | 127.0.0.1:5433 | Replica | streaming |  3 |   0/40004E8 |   0 |  0/40004E8 |   0 |
| postgresql2 | 127.0.0.1:5434 | Replica | streaming |  3 |   0/40004E8 |   0 |  0/40004E8 |   0 |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
2023-09-12 11:52:27.50978 Successfully failed over to "postgresql2"
+ Cluster: batman (7277694203142172922) -+---------+----+-------------+---------+------------+---------+
| Member      | Host           | Role    | State   | TL | Receive LSN |     Lag | Replay LSN |     Lag |
+-------------+----------------+---------+---------+----+-------------+---------+------------+---------+
| postgresql0 | 127.0.0.1:5432 | Replica | stopped |    |     unknown | unknown |    unknown | unknown |
| postgresql1 | 127.0.0.1:5433 | Replica | running |  3 |   0/4000188 |       0 |  0/4000188 |       0 |
| postgresql2 | 127.0.0.1:5434 | Leader  | running |  3 |             |         |            |         |
+-------------+----------------+---------+---------+----+-------------+---------+------------+---------+

patronictl flush

Synopsis

flush
  CLUSTER_NAME
  [ MEMBER_NAME [, ... ] ]
  { restart | switchover }
  [ --group CITUS_GROUP ]
  [ { -r | --role } { leader | primary | standby-leader | replica | standby | any } ]
  [ --force ]

Description

patronictl flush discards scheduled events, if any.

Parameters

CLUSTER_NAME
Name of the Patroni cluster.

MEMBER_NAME
Discard scheduled events for the given Patroni member(s).

Specify one or more members. If omitted, the command considers all members.

Note

Only used if discarding scheduled restart events.

restart
Discard scheduled restart events.

switchover
Discard scheduled switchover event.

--group
Discard scheduled events from the given Citus group.

CITUS_GROUP is the ID of the Citus group.

-r / --role
Discard scheduled events for members that have the given role.

Role can be one of:

leader: the leader of either a regular Patroni cluster or a standby Patroni cluster; or
primary: the leader of a regular Patroni cluster; or
standby-leader: the leader of a standby Patroni cluster; or
replica: a replica of a Patroni cluster; or
standby: same as replica; or
any: any role. Same as omitting this parameter.

Note

Only used if discarding scheduled restart events.

--force
Flag to skip confirmation prompts when performing the flush.

Useful for scripts.

Examples

Discard a scheduled switchover event:

$ patronictl -c postgres0.yml flush batman switchover --force
Success: scheduled switchover deleted

Discard scheduled restart of all standby nodes:

$ patronictl -c postgres0.yml flush batman restart -r replica --force
+ Cluster: batman (7277694203142172922) -+-----------+----+-------------+-----+------------+-----+---------------------------+
| Member      | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag | Scheduled restart         |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+---------------------------+
| postgresql0 | 127.0.0.1:5432 | Leader  | running   |  5 |             |     |            |     | 2025-03-23T18:00:00-03:00 |
| postgresql1 | 127.0.0.1:5433 | Replica | streaming |  5 |   0/4000400 |   0 |  0/4000400 |   0 | 2025-03-23T18:00:00-03:00 |
| postgresql2 | 127.0.0.1:5434 | Replica | streaming |  5 |   0/4000400 |   0 |  0/4000400 |   0 | 2025-03-23T18:00:00-03:00 |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+---------------------------+
Success: flush scheduled restart for member postgresql1
Success: flush scheduled restart for member postgresql2

Discard scheduled restart of nodes postgresql0 and postgresql1:

$ patronictl -c postgres0.yml flush batman postgresql0 postgresql1 restart --force
+ Cluster: batman (7277694203142172922) -+-----------+----+-------------+-----+------------+-----+---------------------------+
| Member      | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag | Scheduled restart         |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+---------------------------+
| postgresql0 | 127.0.0.1:5432 | Leader  | running   |  5 |             |     |            |     | 2025-03-23T18:00:00-03:00 |
| postgresql1 | 127.0.0.1:5433 | Replica | streaming |  5 |   0/4000400 |   0 |  0/4000400 |   0 | 2025-03-23T18:00:00-03:00 |
| postgresql2 | 127.0.0.1:5434 | Replica | streaming |  5 |   0/4000400 |   0 |  0/4000400 |   0 | 2025-03-23T18:00:00-03:00 |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+---------------------------+
Success: flush scheduled restart for member postgresql0
Success: flush scheduled restart for member postgresql1

patronictl history

Synopsis

history
  [ CLUSTER_NAME ]
  [ --group CITUS_GROUP ]
  [ { -f | --format } { pretty | tsv | json | yaml } ]

Description

patronictl history shows a history of failover and switchover events from the cluster, if any.

The following information is included in the output:

TL
Postgres timeline at which the event occurred.

LSN
Postgres LSN at which the event occurred.

Reason
Reason fetched from the Postgres .history file.

Timestamp
Time when the event occurred.

New Leader
Patroni member that has been promoted during the event.

Parameters

CLUSTER_NAME
Name of the Patroni cluster.

If omitted, patronictl uses scope if configured.

--group
Show history of events from the given Citus group.

CITUS_GROUP is the ID of the Citus group.

If omitted, patronictl uses citus.group if configured.

-f / --format
How to format the list of events in the output.

Format can be one of:

pretty: prints history as a pretty table; or
tsv: prints history as tabular information, with columns delimited by \t; or
json: prints history in JSON format; or
yaml: prints history in YAML format.

The default is pretty.

--force
Flag to skip confirmation prompts when performing the flush.

Useful for scripts.

Examples

Show the history of events:

$ patronictl -c postgres0.yml history batman
+----+----------+------------------------------+----------------------------------+-------------+
| TL |      LSN | Reason                       | Timestamp                        | New Leader  |
+----+----------+------------------------------+----------------------------------+-------------+
|  1 | 24392648 | no recovery target specified | 2023-09-11T22:11:27.125527+00:00 | postgresql0 |
|  2 | 50331864 | no recovery target specified | 2023-09-12T11:34:03.148097+00:00 | postgresql0 |
|  3 | 83886704 | no recovery target specified | 2023-09-12T11:52:26.948134+00:00 | postgresql2 |
|  4 | 83887280 | no recovery target specified | 2023-09-12T11:53:09.620136+00:00 | postgresql0 |
+----+----------+------------------------------+----------------------------------+-------------+

Show the history of events in YAML format:

$ patronictl -c postgres0.yml history batman -f yaml
- LSN: 24392648
  New Leader: postgresql0
  Reason: no recovery target specified
  TL: 1
  Timestamp: '2023-09-11T22:11:27.125527+00:00'
- LSN: 50331864
  New Leader: postgresql0
  Reason: no recovery target specified
  TL: 2
  Timestamp: '2023-09-12T11:34:03.148097+00:00'
- LSN: 83886704
  New Leader: postgresql2
  Reason: no recovery target specified
  TL: 3
  Timestamp: '2023-09-12T11:52:26.948134+00:00'
- LSN: 83887280
  New Leader: postgresql0
  Reason: no recovery target specified
  TL: 4
  Timestamp: '2023-09-12T11:53:09.620136+00:00'

patronictl list

Synopsis

list
  [ CLUSTER_NAME [, ... ] ]
  [ --group CITUS_GROUP ]
  [ { -e | --extended } ]
  [ { -t | --timestamp } ]
  [ { -f | --format } { pretty | tsv | json | yaml } ]
  [ { -W | { -w | --watch } TIME } ]

Description

patronictl list shows information about Patroni cluster and its members.

The following information is included in the output:

Cluster
Name of the Patroni cluster.

Member
Name of the Patroni member.

Host
Host where the member is located.

Role
Current role of the member.

Can be one among:

Leader: the current leader of a regular Patroni cluster; or
Standby Leader: the current leader of a Patroni standby cluster; or
Sync Standby: a synchronous standby of a Patroni cluster with synchronous mode enabled; or
Replica: a regular standby of a Patroni cluster.

State
Current state of Postgres in the Patroni member.

Some examples among the possible states:

running: if Postgres is currently up and running;
streaming: if a replica and Postgres is currently streaming WALs from the primary node;
in archive recovery: if a replica and Postgres is currently fetching WALs from the archive;
stopped: if Postgres had been shut down;
crashed: if Postgres has crashed.

TL
Current Postgres timeline in the Patroni member.

Receive LSN
Last WAL location received through streaming replication and synchronized to disk: pg_catalog.pg_last_(xlog|wal)_receive_(location|lsn)().

Receive Lag
Replication lag between the Receive LSN position of the member and its upstream in MB.

Replay LSN
Last WAL location replayed during member recovery: pg_catalog.pg_last_(xlog|wal)_replay_(location|lsn)().

Replay Lag
Replication lag between the Replay LSN position of the member and its upstream in MB.

Besides that, the following information may be included in the output:

System identifier
Postgres system identifier.

Note

Shown in the table header.

Only shown if output format is pretty.

Group
Citus group ID.

Note

Shown in the table header.

Only shown if a Citus cluster.

Pending restart
* indicates that the node needs a restart for some Postgres configuration to take effect. An empty value indicates the node does not require a restart.

Note

Shown as a member attribute.

Shown if:

Printing in pretty or tsv format and with extended output enabled; or
If node requires a restart.

Scheduled restart
Scheduled restart time for the member’s Postgres instance. Empty if no restart is scheduled.

Note

Shown as a member attribute.

Shown if:

Printing in pretty or tsv format and with extended output enabled; or
If node has a scheduled restart.

Tags
Member tags. Empty if no tags are set or all tags have default values.

Note

Shown as a member attribute.

Shown if:

Printing in pretty or tsv format and with extended output enabled; or
If node has any custom tags, or any default tags with non-default values.

Scheduled switchover
Timestamp at which a switchover has been scheduled for the Patroni cluster, if any.

Note

Shown in the table footer.

Only shown if there is a scheduled switchover, and output format is pretty.

Maintenance mode

If the cluster monitoring is currently paused.
Note
Shown in the table footer.
Only shown if the cluster is paused, and output format is pretty.

Parameters

CLUSTER_NAME
Name of the Patroni cluster.

If omitted, patronictl uses scope if configured.

--group
Show information about members from the given Citus group.

CITUS_GROUP is the ID of the Citus group.

-e / --extended
Show extended information.

Force showing Pending restart, Scheduled restart and Tags attributes, even if their value is empty.

Note

Only applies to pretty and tsv output formats.

-t / --timestamp
Print timestamp before printing information about the cluster and its members.

-f / --format
How to format the list of events in the output.

Format can be one of:

pretty: prints history as a pretty table; or
tsv: prints history as tabular information, with columns delimited by \t; or
json: prints history in JSON format; or
yaml: prints history in YAML format.

The default is pretty.

-W
Automatically refresh information every 2 seconds.

-w / --watch
Automatically refresh information at the specified interval.

TIME is the interval between refreshes, in seconds.

Examples

Show information about the cluster in pretty format:

$ patronictl -c postgres0.yml list batman
+ Cluster: batman (7277694203142172922) -+-----------+----+-------------+-----+------------+-----+
| Member      | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
| postgresql0 | 127.0.0.1:5432 | Leader  | running   |  5 |             |     |            |     |
| postgresql1 | 127.0.0.1:5433 | Replica | streaming |  5 |   0/40004E8 |   0 |  0/40004E8 |   0 |
| postgresql2 | 127.0.0.1:5434 | Replica | streaming |  5 |   0/40004E8 |   0 |  0/40004E8 |   0 |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+

Show information about the cluster in pretty format with extended columns:

$ patronictl -c postgres0.yml list batman -e
+ Cluster: batman (7277694203142172922) -+-----------+----+-------------+-----+------------+-----+-----------------+------------------------+-------------------+------+
| Member      | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag | Pending restart | Pending restart reason | Scheduled restart | Tags |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+-----------------+------------------------+-------------------+------+
| postgresql0 | 127.0.0.1:5432 | Leader  | running   |  5 |             |     |            |     |                 |                        |                   |      |
| postgresql1 | 127.0.0.1:5433 | Replica | streaming |  5 |   0/40004E8 |   0 |  0/40004E8 |   0 |                 |                        |                   |      |
| postgresql2 | 127.0.0.1:5434 | Replica | streaming |  5 |   0/40004E8 |   0 |  0/40004E8 |   0 |                 |                        |                   |      |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+-----------------+------------------------+-------------------+------+

Show information about the cluster in YAML format, with timestamp of execution:

$ patronictl -c postgres0.yml list batman -f yaml -t
2023-09-12 13:30:48
- Cluster: batman
  Host: 127.0.0.1:5432
  Member: postgresql0
  Role: Leader
  State: running
  TL: 5
- Cluster: batman
  Host: 127.0.0.1:5433
  Receive LSN: 0/40004E8
  Receive Lag: 0
  Replay LSN: 0/40004E8
  Replay Lag: 0
  Member: postgresql1
  Role: Replica
  State: streaming
  TL: 5
- Cluster: batman
  Host: 127.0.0.1:5434
  Receive LSN: 0/40004E8
  Receive Lag: 0
  Replay LSN: 0/40004E8
  Replay Lag: 0
  Member: postgresql2
  Role: Replica
  State: streaming
  TL: 5

patronictl pause

Synopsis

pause
  [ CLUSTER_NAME ]
  [ --group CITUS_GROUP ]
  [ --wait ]

Description

patronictl pause temporarily puts the Patroni cluster in maintenance mode and disables automatic failover.

Parameters

CLUSTER_NAME
Name of the Patroni cluster.

If omitted, patronictl uses scope if configured.

--group
Pause the given Citus group.

CITUS_GROUP is the ID of the Citus group.

If omitted, patronictl uses citus.group if configured.

--wait
Wait until all Patroni members are paused before returning control to the caller.

Examples

Put the cluster in maintenance mode, and wait until all nodes have been paused:

$ patronictl -c postgres0.yml pause batman --wait
'pause' request sent, waiting until it is recognized by all nodes
Success: cluster management is paused

patronictl promote-cluster

Synopsis

promote-cluster
  [ CLUSTER_NAME ]
  [ --force ]

Description

patronictl promote-cluster converts a standby cluster into a regular Patroni cluster.

The command removes standby_cluster from dynamic configuration and waits until the leader becomes primary. Before the change, it prints the topology and requests confirmation unless --force is set.

Parameters

CLUSTER_NAME: Name of the Patroni cluster.

If omitted, patronictl uses scope if configured.

--force: Flag to skip confirmation prompts when promoting the cluster.

Useful for scripts.

Examples

Promote the standby cluster to run as a regular Patroni cluster:

$ patronictl -c postgres0.yml promote-cluster batman --force

patronictl query

Synopsis

query
  [ CLUSTER_NAME ]
  [ --group CITUS_GROUP ]
  [ { { -r | --role } { leader | primary | standby-leader | replica | standby | any } | { -m | --member } MEMBER_NAME } ]
  [ { -d | --dbname } DBNAME ]
  [ { -U | --username } USERNAME ]
  [ --password ]
  [ --format { pretty | tsv | json | yaml } ]
  [ { { -f | --file } FILE_NAME | { -c | --command } SQL_COMMAND } ]
  [ --delimiter ]
  [ { -W | { -w | --watch } TIME } ]

Description

patronictl query executes a SQL command or script against a member of the Patroni cluster.

Parameters

CLUSTER_NAME
Name of the Patroni cluster.

If omitted, patronictl uses scope if configured.

--group
Query the given Citus group.

CITUS_GROUP is the ID of the Citus group.

-r / --role
Choose a member that has the given role.

Role can be one of:

leader: the leader of either a regular Patroni cluster or a standby Patroni cluster; or
primary: the leader of a regular Patroni cluster; or
standby-leader: the leader of a standby Patroni cluster; or
replica: a replica of a Patroni cluster; or
standby: same as replica; or
any: any role. Same as omitting this parameter.

-m / --member
Choose a member that has the given name.

MEMBER_NAME is the name of the member to be picked.

-d / --dbname
Database to connect and run the query.

DBNAME is the name of the database. If not given, defaults to USERNAME.

-U / --username
User to connect to the database.

USERNAME name of the user. If not given, defaults to the operating system user running patronictl query.

--password
Prompt for the password of the connecting user.

Patroni uses libpq. You can therefore supply a password through ~/.pgpass or PGPASSWORD.

--format
How to format the output of the query.

Format can be one of:

pretty: prints query output as a pretty table; or
tsv: prints query output as tabular information, with columns delimited by \t; or
json: prints query output in JSON format; or
yaml: prints query output in YAML format.

The default is tsv.

-f / --file
Use a file as source of commands to run queries.

FILE_NAME is the path to the source file.

-c / --command
Run the given SQL command in the query.

SQL_COMMAND is the SQL command to be executed.

--delimiter
The delimiter when printing information in tsv format, or \t if omitted.

-W
Automatically re-run the query every 2 seconds.

-w / --watch
Automatically re-run the query at the specified interval.

TIME is the interval between re-runs, in seconds.

Examples

Run a SQL command as postgres user, and ask for its password:

$ patronictl -c postgres0.yml query batman -U postgres --password -c "SELECT now()"
Password:
now
2023-09-12 18:10:53.228084+00:00

Run a SQL command as postgres user, and take password from libpq environment variable:

$ PGPASSWORD=patroni patronictl -c postgres0.yml query batman -U postgres -c "SELECT now()"
now
2023-09-12 18:11:37.639500+00:00

Run a SQL command and print in pretty format every 2 seconds:

$ patronictl -c postgres0.yml query batman -c "SELECT now()" --format pretty -W
+----------------------------------+
| now                              |
+----------------------------------+
| 2023-09-12 18:12:16.716235+00:00 |
+----------------------------------+
+----------------------------------+
| now                              |
+----------------------------------+
| 2023-09-12 18:12:18.732645+00:00 |
+----------------------------------+
+----------------------------------+
| now                              |
+----------------------------------+
| 2023-09-12 18:12:20.750573+00:00 |
+----------------------------------+

Run a SQL command on database test and print the output in YAML format:

$ patronictl -c postgres0.yml query batman -d test -c "SELECT now() AS column_1, 'test' AS column_2" --format yaml
- column_1: 2023-09-12 18:14:22.052060+00:00
  column_2: test

Run a SQL command on member postgresql2:

$ patronictl -c postgres0.yml query batman -m postgresql2 -c "SHOW port"
port
5434

Run a SQL command on any of the standbys:

$ patronictl -c postgres0.yml query batman -r replica -c "SHOW port"
port
5433

patronictl reinit

Synopsis

reinit
  CLUSTER_NAME
  [ MEMBER_NAME [, ... ] ]
  [ --group CITUS_GROUP ]
  [ --wait ]
  [ --force ]
  [ --from-leader ]

Description

patronictl reinit rebuilds a Postgres standby instance managed by a replica member of the Patroni cluster.

Parameters

CLUSTER_NAME
Name of the Patroni cluster.

MEMBER_NAME
Name of the replica member for which the Postgres instance will be rebuilt.

Multiple replica members can be specified. If no members are specified, the command does nothing.

--group
Rebuild a replica member of the given Citus group.

CITUS_GROUP is the ID of the Citus group.

--wait
Wait until the reinitialization of the Postgres standby node(s) is finished.

--force
Flag to skip confirmation prompts when rebuilding Postgres standby instances.

--from-leader
Flag to get basebackup from leader directly.

Useful for scripts.

Examples

Request a rebuild of all replica members of the Patroni cluster and immediately return control to the caller:

$ patronictl -c postgres0.yml reinit batman postgresql1 postgresql2 --force
+ Cluster: batman (7277694203142172922) -+-----------+----+-------------+-----+------------+-----+
| Member      | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
| postgresql0 | 127.0.0.1:5432 | Leader  | running   |  5 |             |     |            |     |
| postgresql1 | 127.0.0.1:5433 | Replica | streaming |  5 |   0/40004E8 |   0 |  0/40004E8 |   0 |
| postgresql2 | 127.0.0.1:5434 | Replica | streaming |  5 |   0/40004E8 |   0 |  0/40004E8 |   0 |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
Success: reinitialize for member postgresql1
Success: reinitialize for member postgresql2

Request a rebuild of postgresql2 and wait for it to complete:

$ patronictl -c postgres0.yml reinit batman postgresql2 --wait --force
+ Cluster: batman (7277694203142172922) -+-----------+----+-------------+-----+------------+-----+
| Member      | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
| postgresql0 | 127.0.0.1:5432 | Leader  | running   |  5 |             |     |            |     |
| postgresql1 | 127.0.0.1:5433 | Replica | streaming |  5 |   0/40004E8 |   0 |  0/40004E8 |   0 |
| postgresql2 | 127.0.0.1:5434 | Replica | streaming |  5 |   0/40004E8 |   0 |  0/40004E8 |   0 |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
Success: reinitialize for member postgresql2
Waiting for reinitialize to complete on: postgresql2
Reinitialize is completed on: postgresql2

Request a rebuild of postgresql2 and get basebackup from leader directly:

$ patronictl -c postgres0.yml reinit batman postgresql2 --from-leader
+ Cluster: batman (7277694203142172922) -+-----------+----+-------------+-----+------------+-----+
| Member      | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
| postgresql0 | 127.0.0.1:5432 | Leader  | running   |  5 |             |     |            |     |
| postgresql1 | 127.0.0.1:5433 | Replica | streaming |  5 |   0/40004E8 |   0 |  0/40004E8 |   0 |
| postgresql2 | 127.0.0.1:5434 | Replica | streaming |  5 |   0/40004E8 |   0 |  0/40004E8 |   0 |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
Success: reinitialize for member postgresql2

patronictl reload

Synopsis

reload
  CLUSTER_NAME
  [ MEMBER_NAME [, ... ] ]
  [ --group CITUS_GROUP ]
  [ { -r | --role } { leader | primary | standby-leader | replica | standby | any } ]
  [ --force ]

Description

patronictl reload requests a reload of local configuration for one or more Patroni members.

It also triggers pg_ctl reload on the managed Postgres instance, even if nothing has changed.

Parameters

CLUSTER_NAME
Name of the Patroni cluster.

MEMBER_NAME
Request a reload of local configuration for the given Patroni member(s).

Specify one or more members. If omitted, the command considers all members.

--group
Request a reload of members of the given Citus group.

CITUS_GROUP is the ID of the Citus group.

-r / --role
Select members that have the given role.

Role can be one of:

leader: the leader of either a regular Patroni cluster or a standby Patroni cluster; or
primary: the leader of a regular Patroni cluster; or
standby-leader: the leader of a standby Patroni cluster; or
replica: a replica of a Patroni cluster; or
standby: same as replica; or
any: any role. Same as omitting this parameter.

--force
Flag to skip confirmation prompts when requesting a reload of the local configuration.

Useful for scripts.

Examples

Request a reload of the local configuration of all members of the Patroni cluster:

$ patronictl -c postgres0.yml reload batman --force
+ Cluster: batman (7277694203142172922) -+-----------+----+-------------+-----+------------+-----+
| Member      | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
| postgresql0 | 127.0.0.1:5432 | Leader  | running   |  5 |             |     |            |     |
| postgresql1 | 127.0.0.1:5433 | Replica | streaming |  5 |   0/40004E8 |   0 |  0/40004E8 |   0 |
| postgresql2 | 127.0.0.1:5434 | Replica | streaming |  5 |   0/40004E8 |   0 |  0/40004E8 |   0 |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
Reload request received for member postgresql0 and will be processed within 10 seconds
Reload request received for member postgresql1 and will be processed within 10 seconds
Reload request received for member postgresql2 and will be processed within 10 seconds

patronictl remove

Synopsis

remove
  CLUSTER_NAME
  [ --group CITUS_GROUP ]
  [ { -f | --format } { pretty | tsv | json | yaml } ]

Description

patronictl remove removes information of the cluster from the DCS.

It is an interactive action.

Warning

This operation will destroy the information of the Patroni cluster from the DCS.

Parameters

CLUSTER_NAME
Name of the Patroni cluster.

--group
Remove information about the Patroni cluster related with the given Citus group.

CITUS_GROUP is the ID of the Citus group.

-f / --format
How to format the list of members in the output when prompting for confirmation.

Format can be one of:

pretty: prints members as a pretty table; or
tsv: prints members as tabular information, with columns delimited by \t; or
json: prints members in JSON format; or
yaml: prints members in YAML format.

The default is pretty.

Examples

Remove information about Patroni cluster batman from the DCS:

$ patronictl -c postgres0.yml remove batman
+ Cluster: batman (7277694203142172922) -+-----------+----+-------------+-----+------------+-----+
| Member      | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
| postgresql0 | 127.0.0.1:5432 | Leader  | running   |  5 |             |     |            |     |
| postgresql1 | 127.0.0.1:5433 | Replica | streaming |  5 |   0/40004E8 |   0 |  0/40004E8 |   0 |
| postgresql2 | 127.0.0.1:5434 | Replica | streaming |  5 |   0/40004E8 |   0 |  0/40004E8 |   0 |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
Please confirm the cluster name to remove: batman
You are about to remove all information in DCS for batman, please type: "Yes I am aware": Yes I am aware
This cluster currently is healthy. Please specify the leader name to continue: postgresql0

patronictl restart

Synopsis

restart
  CLUSTER_NAME
  [ MEMBER_NAME [, ...] ]
  [ --group CITUS_GROUP ]
  [ { -r | --role } { leader | primary | standby-leader | replica | standby | any } ]
  [ --any ]
  [ --pg-version PG_VERSION ]
  [ --pending ]
  [ --timeout TIMEOUT ]
  [ --scheduled TIMESTAMP ]
  [ --force ]

Description

patronictl restart requests a restart of the Postgres instance managed by a member of the Patroni cluster.

The restart can be performed immediately or scheduled for later.

Parameters

CLUSTER_NAME
Name of the Patroni cluster.

--group
Restart the Patroni cluster related with the given Citus group.

CITUS_GROUP is the ID of the Citus group.

-r / --role
Choose members that have the given role.

Role can be one of:

leader: the leader of either a regular Patroni cluster or a standby Patroni cluster; or
primary: the leader of a regular Patroni cluster; or
standby-leader: the leader of a standby Patroni cluster; or
replica: a replica of a Patroni cluster; or
standby: same as replica; or
any: any role. Same as omitting this parameter.

--any
Restart a single random node among the ones which match the given filters.

--pg-version
Select only members which version of the managed Postgres instance is older than the given version.

PG_VERSION is the Postgres version to be compared.

--pending
Select only members which are flagged as Pending restart.

--timeout: abort a restart that exceeds the timeout. If this occurs on the primary, fail over to a replica.

TIMEOUT is the amount of seconds to wait before aborting the restart.

--scheduled
Schedule a restart to occur at the given timestamp.

TIMESTAMP specifies restart time. Use an unambiguous format, preferably with a time zone. Use now for an immediate restart.

--force
Flag to skip confirmation prompts when requesting the restart operations.

Useful for scripts.

Examples

Restart all members of the cluster immediately:

$ patronictl -c postgres0.yml restart batman --force
+ Cluster: batman (7277694203142172922) -+-----------+----+-------------+-----+------------+-----+
| Member      | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
| postgresql0 | 127.0.0.1:5432 | Leader  | running   |  6 |             |     |            |     |
| postgresql1 | 127.0.0.1:5433 | Replica | streaming |  6 |   0/40004E8 |   0 |  0/40004E8 |   0 |
| postgresql2 | 127.0.0.1:5434 | Replica | streaming |  6 |   0/40004E8 |   0 |  0/40004E8 |   0 |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
Success: restart on member postgresql0
Success: restart on member postgresql1
Success: restart on member postgresql2

Restart a random member of the cluster immediately:

$ patronictl -c postgres0.yml restart batman --any --force
+ Cluster: batman (7277694203142172922) -+-----------+----+-------------+-----+------------+-----+
| Member      | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
| postgresql0 | 127.0.0.1:5432 | Leader  | running   |  6 |             |     |            |     |
| postgresql1 | 127.0.0.1:5433 | Replica | streaming |  6 |   0/40004E8 |   0 |  0/40004E8 |   0 |
| postgresql2 | 127.0.0.1:5434 | Replica | streaming |  6 |   0/40004E8 |   0 |  0/40004E8 |   0 |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
Success: restart on member postgresql1

Schedule a restart to occur at 2023-09-13T18:00-03:00:

$ patronictl -c postgres0.yml restart batman --scheduled 2023-09-13T18:00-03:00 --force
+ Cluster: batman (7277694203142172922) -+-----------+----+-------------+-----+------------+-----+
| Member      | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
| postgresql0 | 127.0.0.1:5432 | Leader  | running   |  6 |             |     |            |     |
| postgresql1 | 127.0.0.1:5433 | Replica | streaming |  6 |   0/40004E8 |   0 |  0/40004E8 |   0 |
| postgresql2 | 127.0.0.1:5434 | Replica | streaming |  6 |   0/40004E8 |   0 |  0/40004E8 |   0 |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
Success: restart scheduled on member postgresql0
Success: restart scheduled on member postgresql1
Success: restart scheduled on member postgresql2

patronictl resume

Synopsis

resume
  [ CLUSTER_NAME ]
  [ --group CITUS_GROUP ]
  [ --wait ]

Description

patronictl resume takes the Patroni cluster out of maintenance mode and re-enables automatic failover.

Parameters

CLUSTER_NAME
Name of the Patroni cluster.

If omitted, patronictl uses scope if configured.

--group
Resume the given Citus group.

CITUS_GROUP is the ID of the Citus group.

If omitted, patronictl uses citus.group if configured.

--wait
Wait until all Patroni members are unpaused before returning control to the caller.

Examples

Put the cluster out of maintenance mode:

$ patronictl -c postgres0.yml resume batman --wait
'resume' request sent, waiting until it is recognized by all nodes
Success: cluster management is resumed

patronictl show-config

Synopsis

show-config
  [ CLUSTER_NAME ]
  [ --group CITUS_GROUP ]

Description

patronictl show-config shows the dynamic configuration of the cluster that is stored in the DCS.

Parameters

CLUSTER_NAME
Name of the Patroni cluster.

If omitted, patronictl uses scope if configured.

--group
Show dynamic configuration of the given Citus group.

CITUS_GROUP is the ID of the Citus group.

If omitted, patronictl uses citus.group if configured.

Examples

Show dynamic configuration of cluster batman:

$ patronictl -c postgres0.yml show-config batman
loop_wait: 10
postgresql:
  parameters:
    max_connections: 250
  pg_hba:
  - host replication replicator 127.0.0.1/32 md5
  - host all all 0.0.0.0/0 md5
  use_pg_rewind: true
retry_timeout: 10
ttl: 30

patronictl switchover

Synopsis

switchover
  [ CLUSTER_NAME ]
  [ --group CITUS_GROUP ]
  [ { --leader | --primary } LEADER_NAME ]
  --candidate CANDIDATE_NAME
  [ --force ]

Description

patronictl switchover performs a switchover in the cluster.

Use this command when the cluster is healthy, for example:

There is a leader;
There are synchronous standbys available in a synchronous cluster.

Note

For an unhealthy cluster, consider patronictl failover.

Parameters

CLUSTER_NAME
Name of the Patroni cluster.

If omitted, patronictl uses scope if configured.

--group
Perform a switchover in the given Citus group.

CITUS_GROUP is the ID of the Citus group.

--leader / --primary
Indicate who is the leader to be demoted at switchover time.

LEADER_NAME should match the name of the current leader in the cluster.

--candidate
The node to be promoted on switchover, and take the primary role.

CANDIDATE_NAME is the name of the node to be promoted.

--scheduled
Schedule a switchover to occur at the given timestamp.

TIMESTAMP specifies switchover time. Use an unambiguous format, preferably with a time zone. Use now for an immediate switchover.

--force
Flag to skip confirmation prompts when performing the switchover.

Useful for scripts.

Examples

Switch over with node postgresql2:

$ patronictl -c postgres0.yml switchover batman --leader postgresql0 --candidate postgresql2 --force
Current cluster topology
+ Cluster: batman (7277694203142172922) -+-----------+----+-------------+-----+------------+-----+
| Member      | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
| postgresql0 | 127.0.0.1:5432 | Leader  | running   |  6 |             |     |            |     |
| postgresql1 | 127.0.0.1:5433 | Replica | streaming |  6 |   0/40004E8 |   0 |  0/40004E8 |   0 |
| postgresql2 | 127.0.0.1:5434 | Replica | streaming |  6 |   0/40004E8 |   0 |  0/40004E8 |   0 |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
2023-09-13 14:15:23.07497 Successfully switched over to "postgresql2"
+ Cluster: batman (7277694203142172922) -+---------+----+-------------+---------+------------+---------+
| Member      | Host           | Role    | State   | TL | Receive LSN |     Lag | Replay LSN |     Lag |
+-------------+----------------+---------+---------+----+-------------+---------+------------+---------+
| postgresql0 | 127.0.0.1:5432 | Replica | stopped |    |     unknown | unknown |    unknown | unknown |
| postgresql1 | 127.0.0.1:5433 | Replica | running |  6 |   0/4000188 |       0 |  0/4000188 |       0 |
| postgresql2 | 127.0.0.1:5434 | Leader  | running |  6 |             |         |            |         |
+-------------+----------------+---------+---------+----+-------------+---------+------------+---------+

Schedule a switchover between postgresql0 and postgresql2 to occur at 2023-09-13T18:00:00-03:00:

$ patronictl -c postgres0.yml switchover batman --leader postgresql0 --candidate postgresql2 --scheduled 2023-09-13T18:00-03:00 --force
Current cluster topology
+ Cluster: batman (7277694203142172922) -+-----------+----+-------------+-----+------------+-----+
| Member      | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
| postgresql0 | 127.0.0.1:5432 | Leader  | running   |  8 |             |     |            |     |
| postgresql1 | 127.0.0.1:5433 | Replica | streaming |  8 |   0/40004E8 |   0 |  0/40004E8 |   0 |
| postgresql2 | 127.0.0.1:5434 | Replica | streaming |  8 |   0/40004E8 |   0 |  0/40004E8 |   0 |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
2023-09-13 14:18:11.20661 Switchover scheduled
+ Cluster: batman (7277694203142172922) -+-----------+----+-------------+-----+------------+-----+
| Member      | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
| postgresql0 | 127.0.0.1:5432 | Leader  | running   |  8 |             |     |            |     |
| postgresql1 | 127.0.0.1:5433 | Replica | streaming |  8 |   0/40004E8 |   0 |  0/40004E8 |   0 |
| postgresql2 | 127.0.0.1:5434 | Replica | streaming |  8 |   0/40004E8 |   0 |  0/40004E8 |   0 |
+-------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
Switchover scheduled at: 2023-09-13T18:00:00-03:00
                    from: postgresql0
                    to: postgresql2

patronictl topology

Synopsis

topology
  [ CLUSTER_NAME [, ... ] ]
  [ --group CITUS_GROUP ]
  [ { -W | { -w | --watch } TIME } ]

Description

patronictl topology shows cluster and member information as a tree.

The following information is included in the output:

Cluster
Name of the Patroni cluster.

Note

Shown in the table header.

System identifier
Postgres system identifier.

Note

Shown in the table header.

Member
Name of the Patroni member.

Note

Information in this column is shown as a tree view of members in terms of replication connections.

Host
Host where the member is located.

Role
Current role of the member.

Can be one among:

Leader: the current leader of a regular Patroni cluster; or
Standby Leader: the current leader of a Patroni standby cluster; or
Sync Standby: a synchronous standby of a Patroni cluster with synchronous mode enabled; or
Replica: a regular standby of a Patroni cluster.

State
Current state of Postgres in the Patroni member.

Some examples among the possible states:

running: if Postgres is currently up and running;
streaming: if a replica and Postgres is currently streaming WALs from the primary node;
in archive recovery: if a replica and Postgres is currently fetching WALs from the archive;
stopped: if Postgres had been shut down;
crashed: if Postgres has crashed.

TL
Current Postgres timeline in the Patroni member.

Receive LSN
Last WAL location received through streaming replication and synchronized to disk: pg_catalog.pg_last_(xlog|wal)_receive_(location|lsn)().

Receive Lag
Replication lag between the Receive LSN position of the member and its upstream in MB.

Replay LSN
Last WAL location replayed during member recovery: pg_catalog.pg_last_(xlog|wal)_replay_(location|lsn)().

Replay Lag
Replication lag between the Replay LSN position of the member and its upstream in MB.

Besides that, the following information may be included in the output:

Group
Citus group ID.

Note

Shown in the table header.

Only shown if a Citus cluster.

Pending restart
* indicates the node needs a restart for some Postgres configuration to take effect. An empty value indicates the node does not require a restart.

Note

Shown as a member attribute.

Shown if node requires a restart.

Scheduled restart
Scheduled restart time for the member’s Postgres instance. Empty if no restart is scheduled.

Note

Shown as a member attribute.

Shown if node has a scheduled restart.

Tags
Member tags. Empty if no tags are set or all tags have default values.

Note

Shown as a member attribute.

Shown if node has any custom tags, or any default tags with non-default values.

Scheduled switchover
Timestamp at which a switchover has been scheduled for the Patroni cluster, if any.

Note

Shown in the table footer.

Only shown if there is a scheduled switchover.

Maintenance mode

If the cluster monitoring is currently paused.
Note
Shown in the table footer.
Only shown if the cluster is paused.

Parameters

CLUSTER_NAME
Name of the Patroni cluster.

If omitted, patronictl uses scope if configured.

--group
Show information about members from the given Citus group.

CITUS_GROUP is the ID of the Citus group.

-W
Automatically refresh information every 2 seconds.

-w / --watch
Automatically refresh information at the specified interval.

TIME is the interval between refreshes, in seconds.

Examples

Show topology of the cluster batman – postgresql1 and postgresql2 are replicating from postgresql0:

$ patronictl -c postgres0.yml topology batman
+ Cluster: batman (7277694203142172922) ---+-----------+----+-------------+-----+------------+-----+
| Member        | Host           | Role    | State     | TL | Receive LSN | Lag | Replay LSN | Lag |
+---------------+----------------+---------+-----------+----+-------------+-----+------------+-----+
| postgresql0   | 127.0.0.1:5432 | Leader  | running   |  8 |             |     |            |     |
| + postgresql1 | 127.0.0.1:5433 | Replica | streaming |  8 |   0/40004E8 |   0 |  0/40004E8 |   0 |
| + postgresql2 | 127.0.0.1:5434 | Replica | streaming |  8 |   0/40004E8 |   0 |  0/40004E8 |   0 |
+---------------+----------------+---------+-----------+----+-------------+-----+------------+-----+

patronictl version

Synopsis

version
  [ CLUSTER_NAME [, ... ] ]
  [ MEMBER_NAME [, ... ] ]
  [ --group CITUS_GROUP ]

Description

patronictl version shows the patronictl version. It can also show versions for Patroni clusters and their members.

Parameters

CLUSTER_NAME
Name of the Patroni cluster.

MEMBER_NAME
Name of the member of the Patroni cluster.

--group
Consider a Patroni cluster with the given Citus group.

CITUS_GROUP is the ID of the Citus group.

Examples

Get version of patronictl only:

$ patronictl -c postgres0.yml version
patronictl version 4.0.0

Get version of patronictl and of all members of cluster batman:

$ patronictl -c postgres0.yml version batman
patronictl version 4.0.0

postgresql0: Patroni 4.0.0 PostgreSQL 16.4
postgresql1: Patroni 4.0.0 PostgreSQL 16.4
postgresql2: Patroni 4.0.0 PostgreSQL 16.4

Get version of patronictl and of members postgresql1 and postgresql2 of cluster batman:

$ patronictl -c postgres0.yml version batman postgresql1 postgresql2
patronictl version 4.0.0

postgresql1: Patroni 4.0.0 PostgreSQL 16.4
postgresql2: Patroni 4.0.0 PostgreSQL 16.4