Configuration parameters

This article describes the parameters that can be configured for ADPG services via ADCM. To read about the configuring process, refer to the relevant article Configure services.

NOTE
Some of the parameters become visible in the ADCM UI after the Advanced flag being set.

ADPG

Parameter Description Default value

Data directory

Directories that are used to store data on the ADPG hosts

/pg_data1

Sysctl parameters

This section describes Linux kernel parameters. To add a new parameter, press plus. In the row that appears, specify the parameter name and value in the following format: parameter name:parameter value. You can add multiple parameters by clicking plus.

The example is listed below.

The Sysctl parameters section
The Sysctl parameters section

External Etcd cluster

ADPG uses Etcd as the Patroni Distributed Configuration Store (DCS). If you use an external Etcd cluster, switch on the corresponding toggle button and press plus. In the row that appears, specify Etcd IP address and port in the following format IP address:port. You can add multiple external Etcd servers by clicking plus.

External Etcd settings
External Etcd settings

Patroni ADPG configurations

This section describes Patroni configuration settings that can be changed in ADCM UI. For more information on these settings, refer to ADPG High Availability overview.

Parameter name Description Default value

synchronous_mode

Enables Patroni synchronous mode

false

synchronous_node_count

Specifies the number of synchronous standby nodes. The parameter value should not be more than the count of hosts with the ADPG component of the ADPG service installed on. This parameter is ignored if synchronous_mode is disabled

1

synchronous_mode_strict

Prevents Patroni from switching off the synchronous replication on the primary when no synchronous standby candidates are available. Enables synchronous_mode_strict in addition to synchronous_mode to guarantee that each entry is stored on at least two nodes

false

maximum_lag_on_failover

Specifies the amount of transactions in bytes that can be lost

1048576

ttl

The TTL to acquire the leader lock (in seconds). It defines the length of time before initiation of the automatic failover process

30

loop_wait

The time that Patroni waits (sleeps) before starting a new loop iteration (in seconds)

10

retry_timeout

The timeout for DCS and PostgreSQL operation retries (in seconds). DCS or network issues shorter than this value do not cause Patroni to demote the leader

10

patroni_listen_port

The port that Patroni listens to for the REST API. Can only be changed before the ADPG service installation

8008

ADPG configurations

The ADPG configurations section allows you to specify parameters from the postgresql.conf file (the default path is /pg_data1/adpg14/postgresql.conf). Use ADCM to set these parameters. When the ADPG service executes the Reconfigure & Restart or Reconfigure & Reload action, postgresql.conf is rewritten with settings specified in CLUSTERS → ADPG cluster → Services → ADPG → Configuration. If you make changes directly to postgresql.conf, these changes will be lost.

After modifying ADPG configuration parameters listed below, execute the Reconfigure & Reload or Reconfigure & Restart action to apply changes.

Parameter name Description Default value

port

The TCP port the server listens on. Execute the Reconfigure & Restart action to apply changes

5432

max_connections

Determines the maximum number of concurrent connections to the server. For a replica host, the value of this parameter must be greater than or equal to the value on the leader host. If this requirement is not met, the replica host will reject all requests. Execute the Reconfigure & Restart action to apply changes

100

shared_buffers

Sets the amount of memory the server used for the shared memory buffer. The higher the value of this parameter, the less the load on the host hard drives will be. Execute the Reconfigure & Restart action to apply changes

128 MB

max_worker_processes

Sets the maximum number of background processes that the system can support. Execute the Reconfigure & Restart action to apply changes

8

max_parallel_workers

Sets the maximum number of workers that the system can support for parallel operations. Execute the Reconfigure & Restart action to apply changes

8

max_parallel_workers_per_gather

Sets the maximum number of workers that can be started by a single Gather or Gather Merge node. Execute the Reconfigure & Restart action to apply changes

2

max_parallel_maintenance_workers

Sets the maximum number of parallel workers that can be started by a single utility command. Execute the Reconfigure & Restart action to apply changes

2

effective_cache_size

Tells the ADPG query scheduler what RAM limit is available for caching data for the shared_buffers parameter and the file system cache. The efficie_cache_size parameter helps the scheduler to correctly estimate the costs of efficiently allocating memory for the disk cache per single request. This estimate affects how the cost of using the index will be estimated. The higher this value, the more likely the index scan will be applied. The lower this value, the more likely the sequential scan will be selected

4096 MB

maintenance_work_mem

Sets the memory limit for maintenance operations. Increasing the value for this parameter can lead to faster operations for cleaning and restoring the database from the copy

64 MB

work_mem

Sets the non-shared memory limit that is used for internal operations when the server executes queries (for example, queries to temporary tables, before temporary files on the disk are used). This limit acts as simple resource control, preventing the host from going into swapping due to overallocation. Since the allocated memory is non-shared memory, large and complex requests can exceed the specified limit

4 MB

min_wal_size

Until WAL disk usage stays below the min_wal_size value, old WAL files are recycled for future use at a checkpoint instead of removing. This ensures that enough WAL space is reserved to handle spikes in WAL usage, for example, when running large batch jobs

80 MB

max_wal_size

Sets the memory limit to which the log size can grow between automatic checkpoints. Increasing this setting may increase the recovery time after a failure. The specified limit can be exceeded automatically with a high load on ADPG

1024 MB

wal_keep_size

Sets the minimum size of segments retained in the pg_wal directory, in case a standby server needs to fetch them for streaming replication. If a standby server connected to the sending server falls behind by more than wal_keep_size megabytes, the sending server might remove a WAL segment still needed by the standby. In this case, the replication connection is terminated. Downstream connections also fail as a result. If WAL archiving is enabled, the standby server can fetch the segment from archive and recover. The wal_keep_size parameter sets only the minimum size of segments retained in pg_wal. The system might need to retain more segments for WAL archival or to recover from a checkpoint. If wal_keep_size is zero, the system does not keep any extra segments for standby purposes. In the Enterprise version, if the wal_keep_size value is set to less than 16 MB, Patroni uses 128 MB instead in its configuration. The value in the ADCM UI is not changed

0 MB

huge_pages

Defines whether huge pages can be requested for the main shared memory area. The following values are valid:

  • Try — the server tries to request huge pages. If this operation fails, the server falls back to the default page.

  • On — if a huge page request fails, the server does not start.

  • Off — huge pages are not used.

try

superuser_reserved_connections

Determines the number of connection "slots" that are reserved for PostgreSQL superuser connections. Execute the Reconfigure & Restart action to apply changes

3

logging_collector

Enables the logging collector. The logging collector is a background process that captures log messages sent to stderr and redirects them into log files. Execute the Reconfigure & Restart action to apply changes

The logging collector is enabled (true)

log_directory

Determines the directory that contains log files. It can be specified as an absolute path, or relative to the cluster data directory

log
(the absolute path is /pg_data1/adpg14/log)

log_filename

Specifies the log file name pattern. The value can include strftime %-escapes to define time-varying file names. If you specify a file name pattern without escapes, use a log rotation utility to save disk space

postgresql-%a.log

log_rotation_age

Determines the maximum period of time to use a log file, after which a new log file is created. If this value is specified without units, it is taken as minutes. Set log_rotation_age to 0 to disable time-based log file creation

1d

log_rotation_size

Determines the maximum size of a log file. After a log file reaches the specified size, a new log file is created. If the value is set without units, it is taken as kilobytes. Set log_rotation_size to 0 to disable size-based log file creation

0

log_min_messages

Specifies the minimum severity level of messages that are written to a log file. Valid values are debug5, debug4, debug3, debug2, debug1, info, notice, warning, error, log, fatal, and panic (see Severity levels table). Messages with the specified severity or higher are included in the log file. For example, if you set log_min_messages to warning, the log file will include the warning, error, log, fatal, and panic messages

warning

log_min_error_statement

Specifies which SQL statements that cause errors are logged. Valid values are debug5, debug4, debug3, debug2, debug1, info, notice, warning, error, log, fatal, and panic see Severity levels table). The log file includes SQL statements for messages of the specified severity or higher. To disable error statement logging, set log_min_error_statement to panic

error

postgresql.conf custom section

You can use postgresql.conf custom section to set additional parameters that are written to the postgresql.conf file. To add parameters, switch on the corresponding toggle button and press plus. In the row that appears, specify the parameter name and value in the following format: parameter name:parameter value. You can add multiple parameters by clicking plus.

The postgresql.conf custom section
The postgresql.conf custom section

Parameters from postgresql.conf custom section have lower priority than other parameters. ADCM does not validate these parameters. Changes in the postgresql.conf file is reverted if the Start action of the ADPG service fails.

PG_HBA

The section allows you to add lines to the pg_hba.conf file. The pg_hba.conf file configures the client authentication.

Balancer

Parameter Description Default value

leader as replica

Specifies if the leader handles read only transactions

False

leader_port

Sets a port on the host with the HAProxy component for write transactions

6432

replica_port

Specifies a port on the host with the HAProxy component for read only transactions

6433

balancer_stats_port

Defines a port on which a web page with a HAProxy statistics report is available

7000

Chrony

Parameter Description Default value

NTP servers

Addresses of valid NTP servers. For example, 0.ru.pool.ntp.org, 1.ru.pool.ntp.org. If this parameter is set, then Master takes the time from NTP servers, Secondary Master — from Master or NTP servers, Segment hosts — from Master or Secondary Master. If the parameter is not set or all NTP servers become unavailable, then Master uses its local clocks, Secondary Master uses the Master or local clocks, and Segment Hosts get the time from Master or Secondary Master

 — 

Etcd

IMPORTANT
The Etcd configuration settings cannot be changed after installation.
Parameter Description Default value

listen_peer_urls_port

Specifies the port for peer communication

2380

listen_client_urls_port

Specifies the port for client requests

2379

Metrics storage

Parameter Description Default value

Monitoring db name

Defines a database name that the Metrics storage service should use

adpg_metrics

Monitoring username

Specifies a username to connect to an ADPG monitoring cluster

adpg_metrics_sender

Monitoring user’s password

Specifies a password, which the adpg_metrics_sender user utilizes to connect to an ADPG cluster. The password can include uppercase [A-Z] and lowercase [a-z] English letters, digits [0-9], and special characters

 — 

Grafana TCP port

Defines a TCP port on which Grafana listens for connections

12012

Grafana admin’s password

Specifies a password for the admin user, which is the default username in Grafana

 — 

Found a mistake? Seleсt text and press Ctrl+Enter to report it