: Documentation: 14: shardmanctl : Postgres Professional

shardmanctl
Prev	Up	6.7. Shardman CLI Reference	Home	Next

shardmanctl

shardmanctl — Shardman auxiliary command-line client and deployment tool

Synopsis

shardmanctl [common_options] backup --datadir directory [ --maxtasks number_of_tasks ] --use-ssh

shardmanctl [common_options] daemon check -n | --nodes node_names:port

shardmanctl [common_options] cleanup [ -p | --processrepgroups ] --after-node-operation --after-rebalance

shardmanctl [common_options] config generate [ -f | --file filename]

shardmanctl [common_options] config verify [ -f | --file filename]

shardmanctl [common_options] config get [ -f | --file ] [ -c | --choose-revision ] [ -r | --revision ]

shardmanctl [common_options] config revisions rm [ -r | --revision ] [ -y | --yes ]

shardmanctl [common_options] config rollback [ -r | --revision ] [ -w | --wait time_duration] [ --force ]

shardmanctl [common_options] config revisions [ -f | --format json|text ]

shardmanctl [common_options] config revisions set --keep-config-revisions

shardmanctl [common_options] config update ip [ -u | ip_1=ip_2,hostname_1=hostname_2 ] [ -y | --yes ]

shardmanctl [common_options] config update fdw [ -y | --yes ]

shardmanctl [common_options] cluster repfactor set --value value

shardmanctl [common_options] cluster start

shardmanctl [common_options] cluster stop [ -y | --yes ]

shardmanctl [common_options] cluster topology [ -f | --format table|json|text ]

shardmanctl [common_options] forall --sql query [ --twophase]

shardmanctl [common_options] getconnstr --all

shardmanctl [common_options] init [ -y | --yes ] [ -f | --spec-file spec_file_name] | spec_text

shardmanctl [common_options] load [ -b | --batch-size lines_limit] [ --destination-fields fields_list] [ --distributed-keys key_type_list] [ -D | --delimiter character] [ --null_marker string] [ -e | --escape character] [ -f | --file input_file] [ -F | --format text | csv ] [ -j | --jobs task_total] [ -q | --quote character] [ --reject-file filename] [ --schema filename] [ --source file | postgres ] [ --source-connstr connect_string] [ --source-fields fields_list] [ --source-table table|view|func] [ -t | --table destination_table] [ -h | --help ]

shardmanctl [common_options] nodes add -n | --nodes node_names [ --no-rebalance]

shardmanctl [common_options] nodes start -n | --nodes node_names [ --no-wait]

shardmanctl [common_options] nodes restart -n | --nodes node_names [ --no-wait]

shardmanctl [common_options] nodes stop -n | --nodes node_names [ --no-wait]

shardmanctl [common_options] nodes replace --old old_node --new new_node

shardmanctl [common_options] nodes rm -n | --nodes node_names

shardmanctl [common_options] rebalance [ -f | --force ]

shardmanctl [common_options] recover [ --info file] [ --dumpfile file] [ --shard shard] [ --metadata-only] [ --schema-only] [ --timeout seconds]

shardmanctl [common_options] restart [ -y | --yes ] [ --no-wait ]

shardmanctl [common_options] set pgParam1=value1 [pgParam2=value2 [...]] [ -y | --yes ] [ -w | --wait time_duration] [ -f | --force ]

shardmanctl [common_options] shard -s | --shard shard_name add -n | --node node_names

shardmanctl [common_options] shard -s | --shard shard_name master set -n | --node node_names

shardmanctl [common_options] shard -s | --shard shard_name master reset

shardmanctl [common_options] shard -s | --shard shard_name reset [ -y | --yes ] [ --new-primary | -p ]

shardmanctl [common_options] shard -s | --shard shard_name rm -n | --node node_names [ -f | --force ]

shardmanctl [common_options] shard -s | --shard shard_name switch [ --new-primary node_names]

shardmanctl [common_options] shard -s | --shard shard_name start [ --no-wait ]

shardmanctl [common_options] shard -s | --shard shard_name stop

shardmanctl [common_options] shard -s | --shard shard_name replicas reinit [ --no-wait ] [ -y | --yes ] [ -n | --node node_names]

shardmanctl [common_options] status transactions [ -r | --repgroup replication_group_name]

shardmanctl [common_options] store dump [ -f | --file filename]

shardmanctl [common_options] store keys

shardmanctl [common_options] store lock [ -f | --format text | json ]

shardmanctl [common_options] tables sharded info [ -t | --table table ]

shardmanctl [common_options] tables sharded list

shardmanctl [common_options] tables sharded norebalance

shardmanctl [common_options] tables sharded partmove [ -t | --table table ] [ -s | --shard shard_name ] [ -p | --partnum partition_number ]

shardmanctl [common_options] tables sharded rebalance [ -t | --table table ] [ --skip-run-rebalance ]

shardmanctl [common_options] upgrade

shardmanctl [common_options] bench cleanup

shardmanctl [common_options] bench generate [ -c | --config config_file] [ -o | --output-file file_name]

shardmanctl [common_options] script [ -s | --shard shard_name] [[ -f | --file file_name | --sql query]]

shardmanctl [common_options] psql -s | --shard shard_name

shardmanctl [common_options] history [ -r | --reverse ] [ -f | --format text | json ] [ -l | --limit number_of_commands ]

Here common_options are:

[ --cluster-name cluster_name ] [ --log-level error | warn | info | debug ] [ --monitor-port port] [ --retries retries_number] [ --session-timeout seconds] [ --store-endpoints store_endpoints] [ --store-ca-file store_ca_file] [ --store-cert-file store_cert_file] [ --store-key client_private_key] [ --store-timeout duration] [ --version] [ -h | --help ]

Description

shardmanctl is an utility for managing a Shardman cluster.

For any command that uses the node name as an argument, the node name can be specified either by its hostname or IP address.

The backup command is used to backup a Shardman cluster. A backup consists of a directory with base backups of all replication groups and WAL files needed for recovery. etcd metadata is saved to the etcd_dump file. The backup_info file is created during a backup and contains the backup description. For details of the backup command logic, see Cluster backup with pg_basebackup. For usage details of the command, see the section called “Backing up a Shardman Cluster”.

The cleanup command is used for cleanup after failure of the nodes add command or of the shardmanctl rebalance command. Final changes to the etcd store are done at the end of the command execution. This simplifies the cleanup process. During cleanup, incomplete clover definitions and definitions of the corresponding replication groups are removed from the etcd metadata. Definitions of the corresponding foreign servers are removed from the DBMS metadata of the remaining replication groups. Since the cleanup process can be destructive, by default, the tool operates in the report-only mode: it only shows actions to be done during the actual cleanup. To perform the actual cleanup, add the -p flag. For usage details of the command, see the section called “Performing Cleanup”.

The daemon check command is used to verify that shardmand daemon is running on the nodes specified by --nodes option and is configured for the same cluster as shardmanctl . For usage details of the command, see the section called “Checking shardmand Service on Nodes”.

The init command is used to register a new Shardman cluster in the etcd store or to reinitialize the existing cluster defining a new cluster configuration and removing all data and nodes. In the init mode, shardmanctl reads the cluster specification, processes it and saves to the etcd store as parts of two JSON documents: ClusterSpec — as part of shardman/cluster0/data/cluster and LadleSpec — as part of shardman/cluster0/data/ladle ( cluster0 is the default cluster name used by Shardman utilities). Common options related to the etcd store, such as --store-endpoints, are also saved to the etcd store and pushed down to all Shardman services started by shardmand. For the description of the Shardman initialization file format, see sdmspec.json. For usage details of the command, see the section called “Registering a Shardman Cluster”.

The config generate command is used to create a default sdmspec.json template. By default, data is returned to the standard output. To write the result to a file, use flag -f filename. For the description of the Shardman initialization file format, see sdmspec.json.

The config verify command is used to check a correctness of the input Shardman initialization file. By default, the configuration is read from standard input. To read the configuration from a file, use flag -f filename. For the description of the Shardman initialization file format, see sdmspec.json.

The config get command is used to output the current full cluster specification or a configuration of the specified revision. The command takes the current cluster configuration from the cluster store. For the description of the Shardman initialization file format, see sdmspec.json.

The config update command is used to update the stolon or full Shardman configuration. The new configuration is applied to all replication groups and is saved in shardman/cluster0/data/cluster etcd key. Note that config update can cause a DBMS restart.

The forall command is used to execute an SQL statement on all replication groups in a Shardman cluster.

The getconnstr command is used to get the libpq connection string for connecting to a cluster as administrator.

The load command is used to upload data from a text file to a distributed table or to upload a database schema from a PostgreSQL database to Shardman. When loading data from a file, text and csv formats are supported. If a file is compressed with gzip, it will be automatically decoded while reading. To read data from stdin, specify --file=-. The data loading process can be optimized by specifying the number of parallel workers (key -j).

The nodes add command is used to add new nodes to a Shardman cluster. With the default cross placement policy, nodes are added to a cluster by clovers. Each node in a clover runs the primary DBMS instance and perhaps several replicas of other nodes in the clover. The number of replicas is determined by the Repfactor configuration parameter. So, each clover consists of Repfactor + 1 nodes and can stand loss of Repfactor nodes.

With manual placement policy, each new node is added as a replication group consisting of one primary server. After adding primary nodes, you can add replicas to the new replication group by calling the shard add command.

shardmanctl performs the nodes add operation in several steps:

Acquires a global metadata lock.
For each specified node, checks that shardmand is running on it and that it sees the current cluster configuration.
Calculates the services to be present on each node and saves this information in etcd as part of the shardman/cluster0/data/ladle Layout object.
Generates the configuration for new stolon clusters (also called replication groups) and initializes them.
Registers the added replication groups in the shardman/cluster0/data/ladle etcd key.
Waits for shardmand to start all the necessary services, checks that new replication groups are accessible and have correct configuration.
Creates an auxiliary broadcaster that holds locks on each existing replication group in the cluster.
For each new replication group, copies all schemas and shardman schema data from a randomly selected existing replication group to the new one, ensures that the Shardman extension is installed on the new replication group, and recalculates OIDs used in the extension configuration tables.
On each existing replication group, defines foreign servers referencing the new replication group and recreates definitions of foreign servers on the new replication group.
Recreates all partitions of sharded tables as foreign tables referencing data from old replication groups and has the changes registered in the etcd storage.
For each new replication group, copies the global table data from existing replication groups to the new one.
Rebalances partitions of sharded tables. The rebalancing process for each sharded table iteratively determines the replication group with the maximum and minimum number of partitions and creates a task to move one partition to the replication group with the minimum number of partitions. This process is repeated while max - min > 1. To move partitions, we use logical replication. Partitions of colocated tables are moved together with partitions of the distributed tables to which they refer. You can skip this step using the --no-rebalance.

For usage details of the command, see the section called “Adding Nodes to a Shardman Cluster”.

The nodes rm command is used to remove nodes from a Shardman cluster. In the manual-topology mode, this command only removes the specified nodes from the cluster and if a node is the last in the replication group, the entire group gets removed. In the cross-replication mode, this command removes clovers containing the specified nodes from the cluster. The last clover in the cluster cannot be removed. Any data (such as partitions of sharded relations) on removed replication groups is migrated to the remaining replication groups using logical replication, and all references to the removed replication groups (including definitions of foreign servers) are removed from the metadata of the remaining replication groups. Finally, the metadata in etcd is updated. For usage details of the command, see the section called “Removing Nodes from a Shardman cluster”.

The probackup command is used to backup and restore the Shardman cluster using pg_probackup backup utility. For details of the probackup command logic, see Backup anf Recovery Shardman Backups using pg_probackup. For usage details of the command, see the section called “ probackup ”.

The rebalance command is used to evenly rebalance sharded tables in a cluster. This can be useful, for example, if you did not perform rebalance when adding nodes to the cluster. If the --force option is not provided, then tables with manually moved partitions will be skipped.

The cleanup command with flag --after-rebalance is used to perform cleanup after failure of a rebalance command. On each node, it cleans up subscriptions and publications left from the rebalance command and drops tables that store data of partially-transferred partitions of sharded tables.

The cluster repfactor set command is used to set the value of the replication factor for the Shardman cluster. This command can only be used in manual topology cluster mode. The value of the new replication factor is passed through the command line flag --value repfactor.

The cluster start command is used to start all stopped PostgreSQL instances with the cluster stop command. For the command to work, shardmand must be running.

The cluster stop command is used to stop all PostgreSQL instances for the Shardman cluster. At the same time, the shardmand daemons continue to work.

The cluster topology command is used visualize the topology of a cluster. By default, the topology is returned in a table view. If you want to get a JSON or text representation, then use the flag --format json|text.

The recover command is used to restore a Shardman cluster from a backup created by the backup command. For details of the recover command logic, see Cluster recovery from a backup using pg_basebackup. For usage details of the command, see the section called “Restoring a Shardman Cluster”.

The restart command is used to restart a Shardman cluster, including all shardmand instances. If PostgreSQL instances were previously stopped using the cluster stop command, they will be started. The command returns control after all primary nodes in the cluster have been restarted.

The set command is used to set one or more parameters for DBMS instances of the Shardman cluster. Parameters are passed as arguments to the command line, each of them looks like param=value. The command is actually an alternative to shardmanctl config update -p to update database settings.

The status command is used to display health status of Shardman cluster subsystems. It can show status of several components: store, metadata, shardmand, replication groups, primary nodes and dictionary. If only some subsystems are of interest, option --filter may be used. Also status supports sorting its messages by status, node or replication group and printing the result to stdout as a table (table), text (text) or JSON (json) with table as the default. For usage details of the command, see the section called “Getting the Status of Cluster Subsystems”.

The store dump command gets all the keys and their values from the etcd store and outputs them into the --file, where - value is used for outputting to stdout (default). It is intended to be used for debugging, so some harmless errors may be produced during execution, yet all the available information will be dumped. Only keys for the current cluster (with current cluster prefix like shardman/cluster0/) will be dumped. For usage details of the command, see the section called “Dumping All Keys from the Store to Debug Error Configuration”.

The store get command gets a particular value from the store by its key name. It is expected to be a JSON value, so if it is not (which is not prohibited), some harmless errors may be produced. The key to retrieve from store can be specified with --key option; several keys have aliases — short names for easy use. To get a key by its alias, use --alias option with one of the available aliases (use --help or examples below for reference). Also aliases stolonspec and spec can be used to manipulate initial cluster and stolon configuration explicitly, without retrieving it from the full cluster specification. It is recommended to use existing aliases instead of full key names since there are some additional checks in alias processing, which help to achieve safer results. By default, a key is printed to stdout (explicitly — with --file=- option), but can be output to any desired file. For usage details of the command, see the section called “Getting the Current stolon Specification”.

The store keys command shows all the keys in the store for the current cluster (with cluster prefix) and its aliases. Aliases stolonspec and spec are not shown since they are parts of other keys. For usage details of the command, see the section called “Getting the Cluster and Ladle Key Names For the Current Cluster”.

The store set command creates or rewrites one particular key in the store. It is not expected to be a JSON value for a random key, but if it is one of the keys that have aliases with a known mapping (like ladle or cluster), the command will not accept incorrect JSON structures. Just like store get command, store set accepts a key name via --key or --alias option and the input source file as --file (stdin is specified with - value). For usage details of the command, see the section called “Setting a New Spec for the Cluster”.

The store lock command show the current cluster meta lock information. In case lock does not exist returns Lock not found. Displays cluster id, command that acquired locks, host name and lock time. You can specify --format to output in json format or in text format (by default). For usage details of the command, see the section called “Output Current Cluster Meta Lock Information”.

The upgrade command is used to update the version of Postgresql shardman extension on all cluster nodes. Before upgrading extensions, you need to install new packages and run the restart command. As a result of upgrade, utilities will upgrade shardman and all the other extensions on the server.

Sometimes after running the upgrade command or some user's manual manipulations, dictionary errors may appear in the output of the status command. One of the reasons for these errors is that the value of the srvoptions field of the pg_foreign_server table differs from what the system expects. To solve this specific issue, use the config update fdw command, which will return srvoptions to the expected state.

Note

Most of the described shardmanctl commands take a global metadata lock.

Command-line Reference

This section describes shardmanctl commands. For Shardman common options used by the commands, see the section called “Common Options”.

`backup`

Syntax:

    shardmanctl [common_options] backup --datadir directory [--maxtasks number_of_tasks] [--use-ssh]

Backs up a Shardman cluster.

--datadir directory #

Required.

Specifies the directory to write the output to. If the directory exists, it must be empty. If it does not exist, shardmanctl creates it (but not parent directories).

--maxtasks number_of_tasks #

Specifies the maximum number of concurrent tasks (pg_probackup commands) to run.

Default: number of logical CPUs of the system.

--use-ssh #

If specified shardmanctl recover command will use scp command to restore data. It allows to use backup repository on the local host.

For more details, see the section called “Backing up a Shardman Cluster”

`cleanup`

Syntax:

shardmanctl [common_options] cleanup [-p|--processrepgroups] --after-node-operation|--after-rebalance

Performs cleanup after the nodes add or rebalance command.

-p node_names --processrepgroups=node_names #: Perform an actual cleanup. By default, the tool only shows actions to be done during the actual cleanup. For more details, see the section called “Performing Cleanup”.
--after-node-operation #: Perform cleanup after a failure of a nodes add command.
--after-rebalance #: Perform cleanup after a failure of a rebalance command.

`config update credentials`

Syntax:

shardmanctl [common_options] config update credentials [-u | --user] [-p | --password] [-k | --ssl-key] [-c | --ssl-cert] [-w|--wait time_duration] [--force] [-y | --yes]

Updates password or certificate/key of a user to connect to a Shardman cluster. It only updates the authentication type that was specified by the user (scram-sha-256, ssl) and not the type itself.

-u --user #: User that requires an update of the authentication parameters.
-p --password #: New password.
-k --ssl-key #: New SSL key.
-c --ssl-cert #: New SSL certificate.
-w --wait #: Sets shardmanctl to wait for configuration changes to take effect. If a new configuration cannot be loaded by all replication groups, shardmanctl will wait forever.
--force #: Perform forced update if a cluster operation is in progress.
-y --yes #: Confirm the operation instead of asking approval from the standard input.

`cluster repfactor set`

Syntax:

shardmanctl [common_options] cluster repfactor set --value new_repfactor

Sets the replication factor for the manual-topology mode.

--value=new_repfactor #: New replication factor value

`cluster start`

Syntax:

shardmanctl [common_options] cluster start

Starts all PostgreSQL server instances.

`cluster stop`

Syntax:

shardmanctl [common_options] cluster stop [-y|--yes]

Stops all PostgreSQL server instances.

-y --yes #: confirm the operation instead of asking approval from the standard input.

`cluster topology`

Syntax:

shardmanctl [common_options] cluster topology -f|--format table|json|text

Displays the cluster topology.

-f table|json|text --format=table|json|text #: Output format. For more details, see the section called “Displaying the Cluster Topology”.

`daemon check`

Syntax:

shardmanctl [common_options] daemon check -n|--nodes node_name:port

Checks shardmand on nodes.

-n node_name:port --nodes=node_name:port #: List of nodes to check shardmand on. For more details, see the section called “Checking shardmand Service on Nodes”.

`forall`

Syntax:

shardmanctl [common_options] forall --sql query[ --sql query[ --sql query ...]] [--twophase]

Executes an SQL statement on all replication groups in a Shardman cluster.

--sql query #: Specifies the statement to be executed.
--twophase #: Use the two-phase-commit protocol to execute the statement.

`getconnstr`

Syntax:

shardmanctl [common_options] getconnstr --all

Gets the libpq connection string for connecting to a cluster as administrator.

--all #: Adds replicas to getconnstr.

`init`

Syntax:

    shardmanctl [common_options] init [-y|--yes] [-f|--spec-file spec_file_name]|spec_text

Registers a new Shardman cluster in the etcd store or reinitializes the existing cluster defining a new cluster configuration and removing all data and nodes.

-f spec_file_name --specfile=spec_file_name #: Specifies the file with the cluster specification string. The value of - means the standard input. By default, the string is passed in spec_text. For usage details, see the section called “Registering a Shardman Cluster”.
-y --yes #: Confirm the operation instead of asking approval from the standard input.

`intcheck`

Syntax:

    shardmanctl [common_options] intcheck [-s|--system] [-u|--user] [-c|--catalog] [-o|--output]  [-n|--node node]

Runs pg_integrity_check on all nodes of a Shardman cluster or on a selected one node.

-s --system #: Validate checksums for read-only files. Checksums for read-only files control both file contents and file attributes.
-u --user #: Validate checksums for additional files. Checksums for additional files control both file contents and file attributes.
-c --catalog #: Validate checksums for system catalog tables. For the -c option to work correctly, the database server must be started and accept connections.
-o --output #: Recalculate checksums and write them into a file
-n node_names --node=node_names #: Only execute the pg_integrity_check command on the selected node

`load`

Syntax:

     shardmanctl [common_options] load [ -b | --batch-size lines_limit] [ --destination-fields fields_list]
     [ --distributed-keys key_type_list] [ -D | --delimiter character]
     [--null_marker string] [ -e | --escape character] [ -f | --file input_file]
     [ -F | --format text|csv ] [ -j | --jobs task_total] [ -q | --quote character]
     [ --reject-file filename] [ --schema filename] [ --source file|postgres]
     [ --source-connstr connect_string] [ --source-fields fields_list] [ --source-table source_table]
     [ -t | --table destination_table]

Loads data to a Shardman cluster.

-b lines_limit --batch-size=lines_limit #

Number of rows per batch to write to the Shardman cluster.

Default: 1000.

--destination-fields=fields_list #

Comma-separated list of target table fields. If the value is not set, then all fields of the table are used in the order they are declared.

--distributed-keys=key_type_list #

Comma-separated list of pairs. Each pair consists of a field number (starting with zero) and a type, which are separated by a colon. The following types are supported: bool, char, float4, float8, int2, int4, int8, name, text, varchar and uuid.

-D character --delimiter=character #

Specifies the character that separates columns within each row (line) of the file. This must be a single one-byte character.

Default: tab for text format, comma for CSV format

--null_marker=string #

Specifies the string that represents a null value.

Default: \N for text format, unquoted empty string for CSV format.

-e character --escape=character #

Specifies the character that should appear before a data character that matches the QUOTE value. The default is the same as the QUOTE value (so that the quoting character is doubled if it appears in the data). This must be a single one-byte character. This option is allowed only when using CSV format.

-f filename --file=filename #

Input data filename (or - for stdin)

-F text|csv --format=text|csv #

Input data format. Possible values are text and csv.

Default: text.

-j number --jobs=number #

Number of parallel processes to load data.

Default: number of replication groups.

-q character --quote=character #

Specifies the quoting character to be used when a data value is quoted. The default is double-quote. This must be a single one-byte character. This option is allowed only when using CSV format.

--reject-file=filename #

All data batches with errors during upload will be written to this file. If the value is not set, then such batches will be skipped.

--schema=filename #

The schema that defines the rules for transferring data from PostgreSQL to Shardman. If this option is set, then all other options are not used.

--source=file|postgres #

Data source type — file or postgres.

Default: file.

--source-connstr=string #

Data source database connection string

--source-fields=fields_list #

Comma-separated list of source table fields. If the value is not set, then all fields of the table are used in the order they are declared.

--source-table=table #

Source table, view or function (funcname(param1,...,paramN)).

-t table --table=table #

Destination table.

`nodes add`

Syntax:

shardmanctl [common_options] nodes add -n|--nodes node_names [--no-rebalance]

Adds nodes to a Shardman cluster.

-n node_names --nodes=node_names #

Required.

Specifies the comma-separated list of nodes to be added.

--no-rebalance #

Skip the step of rebalancing partitions of sharded tables. For more details, see the section called “Adding Nodes to a Shardman Cluster”.

`nodes rm`

Syntax:

shardmanctl [common_options] nodes rm -n|--nodes node_names

Removes nodes from a Shardman cluster.

-n node_names --nodes=node_names #: Specifies the comma-separated list of nodes to be removed. For usage details, see the section called “Removing Nodes from a Shardman cluster”.

`probackup`

Syntax:

shardmanctl [common_options] probackup
       [init|archive-command|backup|checkdb|delete|merge|restore|set-config|show|validate]
       [--log-to-console][--help]
       [subcommand_options]

Creates a backup of a Shardman cluster and restores the Shardman cluster from a backup using pg_probackup.

List of subcommands:

init #: Initializes a new repository folder for the Shardman cluster backup and creates a configuration file on all nodes for connection to the backup storage if --storage-type is S3.
archive-command #: Adds archive_command to each replication group (or to a single one if the --shard option is specified) and enables or disables it in the Shardman cluster.
backup #: Creates a backup of the Shardman cluster.
checkdb #: Verifies the Shardman cluster correctness by detecting physical and logical corruption.
delete #: Deletes a backup with a specified ID and the archived WAL files that are no longer in use.
merge #: Merges the backups that belong to a common incremental backup chain. The full backup merges the backups with their first incremental backup. The incremental backup merges the backups with their parent full backup, along with all the incremental backups between them. Once the merge is complete, the full backup covers all the merged data, and the incremental backups are removed as redundant. In this version, you cannot run the merge command using the S3 interface.
restore #: Restores the Shardman cluster from the selected backup.
show #: Shows the list of backups of the Shardman cluster.
validate #: Checks the selected Shardman cluster backup for integrity.
set-config #: Adds the specified settings to the pg_probackup.conf or modifies those previously added.

The following options can be used with all probackup subcommands:

--log-to-console #: Outputs a full probackup log to the console. By default, for each replication group the probackup log file is written to the backup directory (see --backup-path below) as the <backup-directory>/backup/log/pg_probackup-<repgroup-name>.log file. The log rotation file size is 20MB. If this value is reached, the log file is rotated once a shardmanctl probackup validate or shardmanctl probackup backup command is launched.
--help #: Shows subcommand help.

`init`

Syntax:

shardmanctl [common_options] probackup init
-B|--backup-path path
-E|--etcd-path path
[--remote-port port]
[--remote-user username]
[--ssh-key path]
[-t|--timeout seconds]
[-m|--maxtasks number_of_tasks]
[--storage-type mount|remote|S3]
[--s3-config-only]
[--s3-config-path path]
[--s3-host S3_host]
[--s3-port S3_port]
[--s3-access-key S3_access_key]
[--s3-secret-key S3_secret_key]
[--s3-bucket S3_bucket]
[--s3-region S3_region]
[--s3-buffer-size size]
[--s3-retries number_of_retries]
[--s3-timeout time]
[--s3-https]
[-y|--yes]

Initializes a new repository folder for the Shardman cluster backup.

-B path --backup-path path #

Required if --s3-config-only is not used. Specifies the path to the backup catalog where Shardman cluster backups should be stored.

-E path --etcd-path path #

Required if --s3-config-only is not used. Specifies the path to the catalog where the etcd dumps should be stored.

--remote-port port #

Specifies the remote ssh port for replication group instances.

Default: 22.

--remote-user username #

Specifies the remote ssh user for replication group instances.

Default: postgres.

--ssh-key path #

Specifies the ssh private key for execution of remote ssh commands.

Default: $HOME/.ssh/id_rsa.

--storage-type mount|remote|S3 #

Type of the backup storage. If the value is remote, SSH is used to copy data files to the remote backup directory. But this behavior is different if a directory mounted to all nodes or an S3-compatible object storage is used to store backups. To specify these kinds of storage, the value of the --storage-type option is set to mount or S3, respectively.

Default: remote.

--s3-config-path path #

Specifies the path where the S3 configuration file will be created on all Shardman nodes.

Default: <shardman-data-dir>/s3.config.

--s3-config-only #

Create only S3 configuration files on all nodes and skip backup repository initialization. This flag is useful if the value of --storage-type is S3.

--s3-host host #

Specifies the S3 host to connect to S3-compatible storage.

--s3-port port #

Specifies the S3 port to connect to S3-compatible storage.

--s3-access-key access-key #

Specifies the S3 access key to connect to S3-compatible storage.

--s3-secret-key access-key #

Specifies the S3 secret key to connect to the S3-compatible storage.

--s3-bucket bucket #

Specifies the bucket in the S3-compatible object storage for storing backups.

--s3-region bucket #

Specifies the region in the S3-compatible object storage.

--s3-buffer-size size #

Size of the read/write buffer for pg_probackup to communicate with the S3-compatible object storage, in MiB.

Default: 16.

--s3-retries number_of_retries #

Maximum number of attempts for pg_probackup to execute an S3 request in case of failures.

Default: 5.

--s3-timeout time #

Maximum allowable amount of time for pg_probackup to transfer data of size --s3-buffer-size to/from the S3-compatible object storage, in seconds.

Default: 300.

--s3-https #

Specifies the HTTPS URL to connect to the S3-compatible object storage.

-y|--yes #

Approve the operation regardless of whether the file specified in --s3-config-path exists.

`archive-command`

Syntax:

shardmanctl [common_options] probackup archive-command [add|rm]
        -B|--backup-path path
        [-j|--jobs count]
        [--compress]
        [--compress-algorithm algorithm]
        [--compress-level level]
        [--batch-size batch_size]
        [--storage-type mount|remote|S3]
        [--remote-port port]
        [--remote-user username]
        [-s|--shard shard-name]
        [--s3-config-path path]

Adds/removes and enables/disables the archive command for every replication group in the Shardman cluster to put WAL logs into the initialized backup repository.

add #

Adds and enables the archive command for every replication group in the Shardman cluster.

rm #

Disables the archive command in every replication group in the Shardman cluster. No additional options are required.

-B path --backup-path path #

Required when adding archive_command. Specifies the path to the backup catalog where the Shardman cluster backups should be stored.

--batch-size batch_size #

To speed up the archiving, specify the --batch-size option to copy the WAL segments in batches of a specified size. If the --batch-size option is used, it is also possible to specify the -j option to copy a batch of the WAL segments on multiple threads.

--jobs count -j count #

The number of parallel threads that pg_probackup uses when creating a backup. Default: 1.

--compress #

Enables backup compression. If this flag is not specified, compression will be disabled. If the flag is specified, the default zstd algorithm is used with the compression level set to 1, while other compression options are ignored even if they are specified.

--compress-algorithm algorithm #

Defines the compression algorithm: zlib, lz4, zstd, pglz, or none. Once defined, it checks if the values are valid within the scale of the defined algorithm.

The supported compression algorithms depend on the version of Postgres Pro Enterprise that includes the pg_probackup used, as explained in Compression Options.

Default: none.

--compress-level level #

Defines the compression level — 0-9 for zlib, 1 for pglz, 0-22 for zstd, and 0-12 for lz4.

Default: 1.

--storage-type mount|remote|S3 #

Default: remote.

--remote-port port #

Specifies the remote ssh port for replication group instances.

Default: 22.

--remote-user username #

Specifies the remote ssh user for replication group instances.

Default: postgres.

-s|--shard shard-name #

Specifies the name of the shard where the archive command must be added, enabled or disabled. If not specified, the archive command is enabled or disabled for every shard.

--s3-config-path path #

Specifies the path to the S3 configuration file.

Default: <shardman-data-dir>/s3.config.

`backup`

Syntax:

shardmanctl [common_options] probackup backup -B|--backup-path path
        -E|--etcd-path path
        -b|--backup-mode MODE
        [-j|--jobs count]
        [--compress]
        [--compress-algorithm algorithm]
        [--compress-level level]
        [--batch-size batch_size]
        [--storage-type mount|remote|S3]
        [--remote-port port]
        [--remote-user username]
        [--ssh-key path]
        [-t|--timeout seconds]
        [-m|--maxtasks number_of_tasks]
        [--log-directory path]
        [--s3-config-path path]
        [--no-validate]
        [--skip-block-validation]
        [--log-to-console]

Creates a backup of the Shardman cluster.

-B path --backup-path path #

Required. Specifies the path to the backup catalog where Shardman cluster backups should be stored.

-E path --etcd-path path #

Required. Specifies the path to the catalog where the etcd dumps should be stored.

-b MODE --backup-mode MODE #

Required. Defines the backup mode: FULL, PAGE, DELTA, PTRACK.

--batch-size batch_size #

--jobs count -j count #

The number of parallel threads that pg_probackup uses when creating a backup. Default: 1.

--compress #

--compress-algorithm algorithm #

Defines the compression algorithm: zlib, lz4, zstd, pglz, or none.

The supported compression algorithms depend on the version of Postgres Pro Enterprise that includes the pg_probackup used, as explained in Compression Options.

Default: none.

--compress-level level #

Defines the compression level — 0-9 for zlib, 1 for pglz, 0-22 for zstd, and 0-12 for lz4.

Default: 1.

--remote-port port #

Specifies the remote ssh port for replication group instances.

Default: 22.

--remote-user username #

Specifies the remote ssh user for replication group instances.

Default: postgres.

--ssh-key path #

Specifies the ssh private key for execution of remote ssh commands.

Default: $HOME/.ssh/id_rsa.

-t seconds --timeout seconds #

Exit with error after waiting until the cluster is ready for the specified number of seconds.

-m number_of_tasks --maxtasks number_of_tasks #

Specifies the maximum number of concurrent tasks (pg_probackup commands) to run.

Default: number of logical CPUs of the system.

--no-validate #

Skip automatic validation after the backup is taken. You can use this flag if you validate backups regularly and would like to save time when running backup operations.

Default: false.

--skip-block-validation #

Disables block-level checksum verification to speed up the backup process.

Default: false.

--storage-type mount|remote|S3 #

Default: remote.

--log-to-console #

Enables output of the pg_probackup logs to the console.

Default: false.

--log-directory path #

Specifies the directory for pg_probackup logs. Required if --storage-type is set to S3 unless the SDM_LOG_DIRECTORY environment variable is set.

Default: <backup-directory>/backup/log.

--s3-config-path path #

Specifies the path to the S3 configuration file.

Default: <shardman-data-dir>/s3.config.

`checkdb`

Syntax:

shardmanctl [common_options] probackup checkdb
[--amcheck [--skip-block-validation] [--heapallindexed]] [--shard shard]
[-m|--maxtasks number_of_tasks]

Verifies the Shardman cluster correctness by detecting physical and logical corruption.

--amcheck #

Performs logical verification of indexes if no corruption was found while checking data files. You must have the amcheck extension or the amcheck_next extension installed in the database to check its indexes. For databases without amcheck, index verification will be skipped. The amcheck extension is included with the Shardman package.

--heapallindexed #

Checks that all heap tuples that should be indexed are actually indexed. You can use this flag only together with the --amcheck flag. This option is effective depending on the version of amcheck/amcheck_next installed. The amcheck extension included in the Shardman package supports this verification.

--skip-block-validation #

Skip validation of data files. You can use this flag only together with the --amcheck flag, so that only logical verification of indexes is performed.

--shard shard #

Perform the verification only on the specified shard. By default, the verification is performed on all shards.

-m number_of_tasks --maxtasks number_of_tasks #

Specifies the maximum number of concurrent tasks (pg_probackup commands) to run.

Default: number of logical CPUs of the system.

`delete`

Syntax:

shardmanctl [common_options] probackup delete -B|--backup-path path
        -i|--backup-id backup_id
        [-j|--jobs count]
        [-m|--maxtasks number_of_tasks]
        [--storage-type mount|remote|S3]
        [--s3-config-path path]
        [--delete-wal]
        [-y|--yes]

Deletes backup of the Shardman cluster with specified backup_id.

-B path --backup-path path #

Required. Specifies the path to the backup catalog (or key in the bucket of the S3-compatible storage) where Shardman cluster backups should be stored.

-i backup_id --backup-id backup_id #

Required. Specifies the unique identifier of the backup.

--jobs count -j count #

The number of parallel threads that pg_probackup uses when creating a backup. Default: 1.

-m number_of_tasks --maxtasks number_of_tasks #

Specifies the maximum number of concurrent tasks (pg_probackup commands) to run.

Default: number of logical CPUs of the system.

--storage-type mount|remote|S3 #

Default: remote.

To delete the backup that was created with a --storage-type option with a S3 value, set a --storage-type option to a S3 value in the delete command.

--s3-config-path path #

Specifies the path to the S3 configuration file.

Default: <shardman-data-dir>/s3.config.

--delete-wal #

Deletes WAL files that are no longer required to restore the cluster from any of the existing backups.

Default: false.

-y --yes #

Approve operation.

Default: false.

`delete`

Syntax:

shardmanctl [common_options] probackup merge -B|--backup-path path
        -i|--backup-id backup_id
        [-j|--jobs count]
        [-m|--maxtasks number_of_tasks]
        [--no-validate]
        [--no-sync]
        [-y|--yes]

Merges the backups that belong to a common incremental backup chain. The full backup merges the backups with their first incremental backup. The incremental backup merges the backups with their parent full backup, along with all the incremental backups between them. Once the merge is complete, the full backup covers all the merged data, and the incremental backups are removed as redundant.

-B path --backup-path path #

Required. Specifies the path to the backup catalog where Shardman cluster backups should be stored.

-i backup_id --backup-id backup_id #

Required. Specifies the unique identifier of the backup.

--jobs count -j count #

The number of parallel threads that pg_probackup uses when creating a backup. Default: 1.

-m number_of_tasks --maxtasks number_of_tasks #

Specifies the maximum number of concurrent tasks (pg_probackup commands) to run.

Default: number of logical CPUs of the system.

--no-sync #

Do not sync merged files to disk. You can use this flag to speed up the merge process. Using this flag can result in data corruption in case of operating system or hardware crash.

Default: false.

--no-validate #

Skip automatic validation before and after merge.

Default: false.

-y --yes #

Approve the operation.

Default: false.

`restore`

Syntax:

shardmanctl [common_options] probackup restore
        -B|--backup-path path
        -i|--backup-id id
        -j|--jobs count
        [--recovery-target-time timestamp]
        [-I|--recovery-mode incremental_mode]
        [-t|--timeout seconds]
        [-m|--maxtasks number_of_tasks]
        [--metadata-only] [--schema-only] [--shard shard]
        [--no-validate]
        [--skip-block-validation]
        [--s3-config-path path]
        [--storage-type mount|remote|S3]
        [--wal-limit number_of_wal_segments]
        [--log-directory path]

Restores a Shardman cluster from the selected backup.

-B path --backup-path path #

Required. Specifies the path to the backup catalog where Shardman cluster backups should be stored.

-i id --backup-id id #

Required. Specifies backup ID for restore.

--jobs count -j count #

The number of parallel threads that pg_probackup uses when restoring from a backup. Default: 1.

--recovery-target-time timestamp #

Point-in-Time Recovery (PITR) option. Specifies the timestamp for restore. Example: '2024-01-25 15:30:36' in UTC.

-I incremental_mode --recovery-mode incremental_mode #

Specifies the incremental restore mode to be used. Possible values are:

checksum — replace only pages with mismatched checksum and LSN.
lsn — replace only pages with LSN greater than point of divergence.
none — regular restore, default.

-t seconds --timeout seconds #

Exit with error after waiting until the cluster is ready or the recovery is complete for the specified number of seconds.

--metadata-only #

Perform metadata-only restore. By default, full restore is performed.

--schema-only #

Perform schema-only restore. By default, full restore is performed.

--shard shard #

Perform restoring only on the specified shard. By default, restoring is performed on all shards.

--no-validate #

Skip backup validation. You can use this flag if you validate backups regularly and would like to save time when running restore operations.

Default: false.

--skip-block-validation #

Disable block-level checksum verification to speed up validation. During automatic validation before the restore only file-level checksums will be verified.

Default: false.

--s3-config-path path #

Specifies the path to the S3 configuration file.

Default: <shardman-data-dir>/s3.config.

--storage-type mount|remote|S3 #

Default: remote.

--wal-limit number_of_wal_segments #

Specifies the number of WAL segments in which the closest synchronization points will be searched in the case of PITR.

Default: 0 — no limit.

--log-directory path #

Specifies the directory for pg_probackup logs. Required if --storage-type is set to S3 unless the SDM_LOG_DIRECTORY environment variable is set.

Default: <backup-directory>/backup/log.

`show`

Syntax:

shardmanctl [common_options] probackup show
        -B|--backup-path path
        [-f|--format table|json]
        [--archive ]
        [-i|--backup-id backup-id]
        [--instance instance]
        [--storage-type mount|remote|S3]
        [--s3-config-path path]

Shows the list of backups of the Shardman cluster.

-B path --backup-path path #

Required. Specifies the path to the backup catalog where Shardman cluster backups should be stored.

-f table|json --format table|json #

Specifies the output format.

Default: table.

--archive #

Shows the WAL archive information.

-i backup-id --backup-idbackup-id #

Shows information about the specific backups.

--instanceinstance #

Shows information about the specific instance.

--s3-config-path path #

Specifies the path to the S3 configuration file.

Default: <shardman-data-dir>/s3.config.

--storage-type mount|remote|S3 #

Default: remote.

`validate`

Syntax:

shardmanctl [common_options] probackup validate
        -B|--backup-path path
        -i|--backup-id id
        [-t|--timeout seconds]
        [-m|--maxtasks number_of_tasks]
        [--log-to-console]
        [--storage-type mount|remote|S3]
        [--s3-config-path path]
        [--log-directory path]

Checks the selected Shardman cluster backup for integrity.

-B path --backup-path path #

Required. Specifies the path to the backup catalog where Shardman cluster backups should be stored.

-i id --backup-id id #

Required. Specifies backup ID for validation.

--log-to-console #

Enables output of pg_probackup logs to the console.

Default: false.

-t seconds --timeout seconds #

Exit with error after waiting until the cluster is ready for the specified number of seconds.

-m number_of_tasks --maxtasks number_of_tasks #

Specifies the maximum number of concurrent tasks (pg_probackup commands) to run.

Default: number of logical CPUs of the system.

--s3-config-path path #

Specifies the path to the S3 configuration file.

Default: <shardman-data-dir>/s3.config.

--storage-type mount|remote|S3 #

Default: remote.

--log-directory path #

Specifies the directory for pg_probackup logs. Required if --storage-type is set to S3 unless the SDM_LOG_DIRECTORY environment variable is set.

Default: <backup-directory>/backup/log.

`set-config`

Syntax:

shardmanctl [common_options] probackup set-config
        [--archive-timeout int]
        [-B | --backup-path string]
        [-m |--maxtasks int]
        [--remote-port int]
        [--remote-user string]
        [--retention-redundancy int]
        [--retention-window int]
        [--wal-depth int]
        [--s3-config-path string]
        [-s |--shard string]
        [--storage-type string]

Adds the specified settings to the pg_probackup.conf or modifies those previously added.

--archive-timeout int #

Sets a timeout for the WAL segment archiving and streaming, in seconds.

Default: pg_probackup waits for 300 seconds.

-B string --backup-path=string #

Specifies the absolute path to the backup catalog.

Default: current value of the pg_probackup.conf.

-m int --maxtasks=int #

Specifies the maximum number of concurrent tasks (pg_probackup commands) to run.

Default: number of logical CPUs of the system.

--remote-port int #

An SSH remote backup port.

Default: current value of the pg_probackup.conf.

--remote-user string #

An SSH remote backup user.

Default: current value of the pg_probackup.conf.

--retention-redundancy int #

Specifies the number of the full backup copies to store in the data directory. It must be set to a non-negative integer. The zero value disables this setting.

Default: current value of the pg_probackup.conf.

--retention-window int #

A number of days of recoverability. It must be set to a non-negative integer. The zero value disables this setting.

Default: current value of the pg_probackup.conf.

--wal-depth int #

A number of the latest valid backups on every timeline that must retain the ability to perform PITR. Must be set to a non-negative integer. The zero value disables this setting.

Default: current value of the pg_probackup.conf.

--s3-config-path string #

A path to the S3 configuration file.

Default: /var/lib/pgpro/sdm-14/data/s3.config

-s string --shard=string #

A name of the shard to make the set-config command for. If not specified, the command is run for all the shards.

Default: current value of the pg_probackup.conf.

--storage-type string #

A backup storage type, the possible values are remote, mount, S3.

Default: remote.

`rebalance`

Syntax:

shardmanctl [common_options] rebalance [-f|--force]

Rebalances sharded tables.

-f --force #: Perform forced rebalance of sharded tables whose partitions were manually moved.

`recover`

Syntax:

shardmanctl [common_options] recover [--info file] [--dumpfile file] [--shard shard] [--metadata-only][--schema-only] [--timeout seconds]

Restores a Shardman cluster from a backup created by the backup command.

--dumpfile file #

Required for metadata-only restore.

Specifies the file to load the etcd metadata dump from.

--info file #

Required for full restore.

Specifies the file to load information about the backup from.

--shard shard #

Perform restoring only on the specified shard. By default, restoring is performed on all shards.

--metadata-only #

Perform metadata-only restore. By default, full restore is performed.

--schema-only #

Perform schema-only restore. By default, full restore is performed.

--timeout seconds #

Exit with error after waiting until the cluster is ready or the recovery is complete for the specified number of seconds.

For more details, see the section called “Restoring a Shardman Cluster”

`restart`

Syntax:

     shardmanctl [common_options] restart [-y|--yes] [--no-wait]

Restarts a Shardman cluster.

-y --yes #: Confirm the operation instead of asking approval from the standard input.
--no-wait #: Do not wait for the replicas to start.

`shard add`

Syntax:

shardmanctl [common_options] shard -s|--shard shard_name add -n|--nodes node_names [--no-wait]

Adds a replica to a shard.

-s shard_name --shard=shard_name #: Shard name.
-n node_names --nodes=node_names #: Specifies the comma-separated list of replica nodes to be added.
--no-wait #: Do not wait for the shard to start.

`shard master set`

Syntax:

shardmanctl [common_options] shard -s|--shard shard_name master set -n| node node_names

Sets the precedence for a certain primary server for a specified shard.

-s shard_name --shard=shard_name #: Shard name.
master set #: Primary server with precedence.
-n node_names --nodes=node_names #: Specifies the comma-separated list of replica nodes.

`shard master reset`

Syntax:

shardmanctl [common_options] shard -s|--shard shard_name master reset

Resets the parameters of the master with precedence for the shard.

-s shard_name --shard=shard_name #: Shard name.
master reset #: Resets the parameters of the master with precedence for the shard.
-n node_names --nodes=node_names #: Specifies the comma-separated list of replica nodes.

`shard add`

Syntax:

shardmanctl [common_options] shard -s|--shard shard_name reset [--yes | -y][--new-primary | -p]

Resets nodes of a replication group if they are in a state of hanging.

-s shard_name --shard=shard_name #: Shard name.
-y --yes #: Confirm the operation instead of asking approval from the standard input.
--new-primary -p #: New primary node host.

`shard rm`

Syntax:

     shardmanctl [common_options] shard -s|--shard shard_name rm -n|--nodes node_names [-f|--force]

Removes a replica from a shard.

-s shard_name --shard=shard_name #: Shard name
-n node_names --nodes=node_names #: Specifies the comma-separated list of replica nodes to be removed.
-f --force #: Perform forced removal of the node, even if it is dead.

`shard switch`

Syntax:

     shardmanctl [common_options] shard -s|--shard shard_name switch [--new-primary node_names]

Switches the primary node.

-s shard_name --shard=shard_name #: Shard name.
--new-primary=node_names #: New primary node host.

`shard start`

Syntax:

     shardmanctl [common_options] shard -s |--shard shard_name start [--no-wait]

Starts the shard.

-s shard_name --shard=shard_name: Shard name.
--no-wait: Do not wait for the shard to start.

`shard stop`

Syntax:

     shardmanctl [common_options] shard -s |--shard shard_name stop

Stops the shard.

-s shard_name --shard=shard_name: Shard name.

`shard replicas reinit`

Syntax:

     shardmanctl [common_options] shard -s|--shard shard_name replicas reinit [-n|--node node_names] [-y|--yes] [--no-wait]

Resets replicas of a specific shard.

-s shard_name --shard=shard_name #: Shard name.
-n node_names --node=node_names #: Specifies the node on which to reset replicas. If not specified, checks shard replicas on all nodes.
-y --yes #: Confirm the operation instead of asking approval from the standard input.
--no-wait #: Do not wait wait for replicas to become ready.

For more details, see the section called “Reinitializing Replicas”

`nodes start`

Syntax:

     shardmanctl [common_options] nodes start -n|--nodes node_names [--no-wait]

Starts the nodes.

-n node_names --nodes=node_names: Node names.
--no-wait: Sets shardmanctl not to wait for the nodes to start.

`nodes restart`

Syntax:

     shardmanctl [common_options] nodes restart -n|--nodes node_names [--no-wait]

Restarts the nodes.

-n node_names --nodes=node_names: Node names.
--no-wait: Do not wait for the nodes to restart.

`nodes stop`

Syntax:

     shardmanctl [common_options] nodes stop -n|--nodes node_names [--no-wait]

Stops the nodes.

-n node_names --nodes=node_names: Node names.
--no-wait: Do not wait for the nodes to stop.

`status`

Syntax:

     shardmanctl [common_options] status [-f|--format table|json] [--filter store|metadata|shardmand|rg|master|dictionary|all] [-s|--sort node|rg|status]

Reports on the health status of Shardman cluster subsystems.

-f table|json --format=table|json #

Specifies the report format.

Default: table.

For more details, see the section called “Getting the Status of Cluster Subsystems”.

Specifies subsystems whose status information should be included in the output.

Default: all.

For more details, see the section called “Getting the Status of Cluster Subsystems”.

-s node|rg|status --sort node|rg|status #

Sort messages inside one group (table) as specified.

Default: node.

For more details, see the section called “Getting the Status of Cluster Subsystems”.

`status transactions`

Syntax:

     shardmanctl [common_options] status transactions [-r|--repgroup replication_group_name

Shows distributed transactions that Shardman built-in monitoring tools failed to resolve.

-r replication_group_name --repgroup=replication_group_name #

Specifies the replication group for which to output transactions.

Default: all replication groups.

For more details, see the section called “Outputting the List of Unresolved Distributed Transactions”.

`store dump`

Syntax:

     shardmanctl [common_options] store dump [-f|--file filename]

Dumps current cluster specifications from the store.

-f filename --file=filename #

Specifies the output file (- for stdout).

Default: -.

For more details, see the section called “Dumping All Keys from the Store to Debug Error Configuration”.

`store lock`

Syntax:

     shardmanctl [common_options] store lock [-f|--format text|json]

Shows the current cluster meta lock information.

-f=text|json --format=text|json #

Specifies the output format.

Default: text.

For more details, see the section called “Output Current Cluster Meta Lock Information”.

`store get`

Syntax:

    shardmanctl [common_options] store get [[-a|--alias aliasname]|[-k|--key keyname] [-f|--file filename]]

Gets the specified key from the store.

-a aliasname --alias=ladle|cluster |spec|stolonspec #

Specifies the use of alias instead of the full key name. Cannot be used with --key.

For more details, see the section called “Getting the Current stolon Specification”.

-k keyname --key=keyname #

Specifies the key to retrieve from the store. Cannot be used with --alias.

For more details, see the section called “Getting the Current stolon Specification”.

-f filename --file=filename #

Specifies the file to print the value to.

Default: - (stdout).

For more details, see the section called “Getting the Current stolon Specification”.

`store keys`

Syntax:

    shardmanctl [common_options] store keys

Gets all keys with the current cluster prefix from the store.

For more details, see the section called “Getting the Cluster and Ladle Key Names For the Current Cluster”.

`store set`

Syntax:

    shardmanctl [common_options] store set [[-a|--alias aliasname]|[-k|--key keyname]] [-f|--file filename]

Creates or rewrites a key in the store.

Specifies the use of alias instead of the full key name. Cannot be used with --key.

-k keyname --key=keyname #

Specifies the key name to set in the store. Cannot be used with --alias.

-f filename --file=filename #

Specifies the file with input data (- for stdin).

For more details, see the section called “Setting a New Spec for the Cluster”.

`tables sharded info`

Syntax:

    shardmanctl [common_options] tables sharded info [-t|--table table_name]

Gets information about a sharded table.

-t table --table=table #: Specifies the name of the table in the format schema.table

`tables sharded list`

Syntax:

    shardmanctl [common_options] tables sharded list

Gets the list of all sharded tables.

`tables sharded norebalance`

Syntax:

    shardmanctl [common_options] tables sharded norebalance

Gets the list of sharded tables with automatic rebalancing disabled.

`tables sharded partmove`

Syntax:

    shardmanctl [common_options] tables sharded partmove [-t|--table table_name] [-s|--shard shard_name] [-p|--partnum number]

Moves the specified partition of a sharded table to a new shard.

-t table --table=table #: Specifies the name of the table in the format schema.table.
-p number --partnum=number #: Specifies the number of the partition to move.
-s shard_name --shard=shard_name #: Specifies the name of the new shard for the partition.

`tables sharded rebalance`

Syntax:

    shardmanctl [common_options] tables sharded rebalance [-t|--table table_name]

Enables and runs automatic data rebalancing for the selected sharded table.

-t table --table=table #: Specifies the name of the table in the format schema.table.

`config get`

Syntax:

shardmanctl [common_options] config get [-c | --choose-revision] [-r | --revision ] [-f | --file]

Outputs the current full cluster specification or a configuration of the specified revision.

-c --choose-revision #: Enables an interactive mode of choosing a configuration of the specified revision.
-r --revision #: ID of a configuration revision.
-f file_name --file=file_name #: Name of a file for writing the configuration. If not specified, the value is stdout.

`config revisions rm`

Syntax:

shardmanctl [common_options] config revisions rm [-r | --revision ] [-y | --yes]

Deletes a specified configuration revision from history.

-r --revision #: ID of a configuration revision. If not specified, enables an interactive mode of choosing a configuration of the specified revision. This is a timestamp of an operation that resulted in Shardman configuration change.
-y --yes #: Perform automatic confirmation.

`config update`

Syntax:

shardmanctl [common_options] config update [[-f|--file stolon_spec_file|shardman_spec_file]|spec_text [-p|--patch][-w|--wait]] [--force]

Updates the stolon or full Shardman configuration.

-f stolon_spec_file|shardman_spec_file --specfile=stolon_spec_file|shardman_spec_file #: Specifies the file with the stolon or full Shardman configuration. The configuration file type is determined automatically. The value of - means the standard input. By default, the configuration is passed in spec_text.
-w --wait #: Sets shardmanctl to wait for configuration changes to take effect. If a new configuration cannot be loaded by all replication groups, shardmanctl will wait forever.
-p --patch #: Merge the new configuration into the existing one. By default, the new configuration replaces the existing one.
--force #: Perform forced update if a cluster operation is in progress.

`config rollback`

Syntax:

shardmanctl [common_options] config rollback [-r | --revision] [-w|--wait time_duration] [--force] [-y|--yes]

Makes a rollback of Shardman to one of the previous states. When rolling back to the config revision that has max_connections, max_prepared_transactions, or max_worker_processes parameters, the replicas are reinitialized.

-r --revision #

ID of a revision the rollback must be made to. It is a timestamp of an operation that resulted in Shardman configuration change.

If not specified, a user is presented with a list of revisions that he can choose from.

-w --wait #

Sets shardmanctl to wait for configuration changes to take effect. If a new configuration cannot be loaded by all replication groups, shardmanctl will wait forever.

Default: 1h.

-f --force #

Perform forced setting of a parameter if a cluster operation is in progress.

-y --yes #

Perform automatic confirmation.

`config revisions`

Syntax:

shardmanctl [common_options] config revisions [-f|--format text|json]

Outputs the revision history of the Shardman cluster configuration. It has the following information for each revision:

revision_id — timestamp of the command that resulted in the Shardman cluster configuration change
host — name of the host from which this command was executed
user — user who executed this command
command — the command itself

-f=text|json --format=text|json #

Specifies the output format.

Default: text.

`config revisions set`

Syntax:

shardmanctl [common_options] config revisions set [--keep-config-revisions]

Allows setting the length of the configuration revision history. This length cannot be lower than 5, in which case it is automatically set to 5. For Shardman clusters where the configuration revision history was not collected yet, the length is automatically set to 20.

--keep-config-revisions #

A limit on the number of revisions for one Shardman configuration. If the limit is lower than the current history length, the older versions out of this limit will be deleted. Also, if the number of operations resulting in configuration changes exceeds the limit, the oldest revision is deleted.

Default: 20.

`config update ip`

Syntax:

shardmanctl [common_options] config update ip [-u|ip_1=ip_2,hostname_1=hostname_2][-y|--yes]

Updates the specified node IPs in the cluster.

-u ip_1=ip_2,hostname_1=hostname_2 #: Specifies the node IPs to be updated.
-y --yes #: Perform automatic confirmation.

`set`

Syntax:

shardmanctl [common_options] set pgParam1=value1 [pgParam2=value2 [...]] [-y|--yes] [-w|--wait time_duration] [-f|--force]

Sets the values of the specified Shardman cluster database parameters.

-w --wait #

Sets shardmanctl to wait for configuration changes to take effect. Value examples: 2h45m, 1m30s, 5m, 10s.

Default: 1h.

-y --yes #

Perform automatic confirmation.

-f --force #

Perform forced setting of a parameter if a cluster operation is in progress.

`upgrade`

Syntax:

shardmanctl [common_options] upgrade

Upgrades the shardman database extension and updates pg_foreign_server options.

`bench init`

Syntax:

shardmanctl [common_options] bench init [--schema-type single|simple|shardman|custom]
[--schema-file file_name] [-s|--scale scale_value]  [-n|--no-vacuum]
[-F|--fillfactor fillfactor_value]

Initializes the benchmark schema via pgbench. Schema can be custom or predefined. Creates tpc-b schema tables and fills them.

--schema-type=single|simple|shardman|custom #

Type of schema used by schema initialization. Possible values:

single — schema for a single PostgreSQL benchmark test
simple — simple sharded schema
shardman — sharded schema optimized for Shardman
custom — schema initialized by the user from the --schema-file file

Default schema: shardman.

--schema-file=file_name #

File with DDL query for the custom schema type, to be used to create tpc-b tables for pgbench: pgbench_accounts, pgbench_branches, pgbench_tellers, pgbench_history.

-s scale_value --scale=scale_value #

Multiply the number of generated rows by the given scale factor.

-n --no-vacuum #

Perform no vacuuming during initialization.

-F fillfactor_value --fillfactor=fillfactor_value #

Fill pgbench tables with the given fillfactor value.

`bench run`

Syntax:

shardmanctl [common_options] bench run [--schema-type single|simple|shardman|custom]
[-f|--file file_name] [-c|--client client_value]  [-C|--connect] [--full-output]
[-j|--jobs jobs_value][-T|--time seconds][-t|--transactions transactions_value]
[-s|--scale scale_factor] [ -P | --progress seconds] [ -R | --rate rate] [ -M | --protocol querymode]

Runs the initialized benchmark via pgbench. Can use the default pgbench script or a custom script from a file.

--schema-type=single|simple|shardman|custom #

Type of schema used by schema initialization (bench init). Possible values:

single — schema for single PostgreSQL benchmark
simple — simple sharded schema
shardman — sharded schema optimized for Shardman
custom — schema initialized by the user from the --schema-file file.

Default schema: shardman.

-f file_name --file=file_name #

Add a transaction script read from filename to the list of scripts to be executed.

Optionally, write an integer weight after @ to adjust the probability of selecting this script versus other ones. The default weight is 1. (To use a script file name that includes an @ character, append a weight so that there is no ambiguity, for example filen@me@1).

-c client_value --client=client_value #

Number of clients simulated, that is, number of concurrent database sessions.

-C --connect #

Establish a new connection for each transaction rather than doing it just once per client session.

--full-output #

Print all pgbench output.

-j jobs_value --jobs=jobs_value #

Number of worker threads within pgbench.

-s scale_factor --scale=scale_factor #

Multiply the number of generated rows by + the given scale factor.

-T seconds --time=seconds #

Run the test for this many seconds instead of a fixed number of transactions per client.

-t transactions_value --transactions=transactions_value #

Number of transactions each client runs.

Default: 10.

-P seconds --progress=seconds #

Show progress report every sec seconds. The report includes the time since the beginning of the run, the TPS since the last report, and the transaction latency average, standard deviation, and the number of failed transactions since the last report. Under throttling (-R), the latency is computed with respect to the transaction scheduled start time, not the actual transaction beginning time, thus it also includes the average schedule lag time. When --max-tries is used to enable transaction retries after serialization/deadlock errors, the report includes the number of retried transactions and the sum of all retries.

-R rate --rate=rate #

Execute transactions targeting the specified rate instead of running as fast as possible (the default). The rate is given in transactions per second. If the targeted rate is above the maximum possible rate, the rate limit won't impact the results.

-M querymode --protocol=querymode #

Protocol to use for submitting queries to the server:

simple: use simple query protocol.
extended: use extended query protocol.
prepared: use extended query protocol with prepared statements.

In the prepared mode, pgbench reuses the parse analysis result starting from the second query iteration, so pgbench runs faster than in other modes.

Default: simple.

`bench cleanup`

Syntax:

shardmanctl [common_options] bench cleanup

Cleans up schema database after benchmarks. Drops tpc-b tables.

`bench generate`

Syntax:

shardmanctl [common_options] bench generate [-c|--config file_name] [-o|--output-file file_name]

Gets the benchmark configuration from a file and generates a bash script to create a schema optimized for Shardman and run the benchmark using pgbench. The configuration file must be in yaml format.

-f file_name --file=file_name #

The configuration file path. The file contains a sequence of script confugurations. Each script must have a schema_type: single|simple|shardman|custom. For a custom schema it is necessary to specify the schema_file with the DDL script. Optional parameters: init_flags (default set: -s 1000), run_flags (default set: -n -P 10 -c 10 -j 4 -T 60), partitions (default value: 50). It is highly recomended to use -n (--no-vacuum) parameter inside run_flags. Configuration file example:

    benches:
    - schema_type: single
      init_flags: "-s 3"
      run_flags: "-n -P 10 -c 10 -j 4 -T 10"
    - schema_type: simple
      init_flags: "-s 4"
      run_flags: "-n -P 10 -c 20 -j 4 -T 10"
      partitions: 100
    - schema_type: shardman
      init_flags: "-s 5"
      run_flags: "-n -P 10 -c 20 -j 4 -T 10"
    - schema_type: custom
      init_flags: "-s 6"
      schema_file: "schema.psql"

-o file_name --output-file=file_name #

Output file. Default: stdout.

`script`

Syntax:

shardmanctl [common_options] script -s|--shard shard_name][[-f|--file file_name][--sql query]]

Executes non-transactional commands from a file or from the command-line on the specified shards.

-s shard_name --shard=shard_name #: Shard name.
-f file_name --file=file_name #: Add a transaction script from the file_name file to the list of scripts to be executed.
--sql query #: Specifies the statement to be executed and can only be used separately from -f.

`psql`

Syntax:

shardmanctl [common_options] psql -s|--shard shard_name

Connects to the first available primary node if no options are specified.

-s shard_name --shard=shard_name #: Name of the shard. If specified, the connection is installed with this shard current primary.

`daemon set`

Syntax:

shardmanctl [common_options] daemon set [--session-log-level debug | info | warn | error] [--session-log-format json|text] [--session-log-nodes]

Allows updating the log parameters “on the fly”.

--session-log-level debug | info | warn | error #

Updates the log level to debug, info, warn, or error.

--session-log-format json|text #

Updates the log output format to text or json.

--session-log-nodes #

Specifies which cluster nodes must be updated. If not specified, the parameters are updated on every node.

Default: all nodes.

`history`

Syntax:

shardmanctl [common_options] history [--reverse | -r] [-f|--format json|text] [-l|--limit number_of_commands]

Shows history of the commands that updated the cluster. By default, they are sorted from the most recent to the oldest ones.

-r --reverse #

Switches to the ascending sorting order.

-f json|text --format=json|text #

Output format.

Default: text.

-l --limit=number_of_commands #

Limit for the number of the most recent commands in the output. The maximum value is 200.

Default: 20.

Common Options

shardmanctl common options are optional parameters that are not specific to the utility. They specify etcd connection settings, cluster name and a few more settings. By default shardmanctl tries to connect to the etcd store 127.0.0.1:2379 and use the cluster0 cluster name. The default log level is info.

-h, --help #: Show brief usage information.
--cluster-name cluster_name #: Specifies the name for a cluster to operate on. The default is cluster0.
--log-level level #: Specifies the log verbosity. Possible values of level are (from minimum to maximum): error, warn, info and debug. The default is info.
--retries number #: Specifies how many times shardmanctl retries a failing etcd request. If an etcd request fails, most likely, due to a connectivity issue, shardmanctl retries it the specified number of times before reporting an error. The default is 5.
--session-timeout seconds #: Specifies the session timeout for shardmanctl locks. If there is no connectivity between shardmanctl and the etcd store for the specified number of seconds, the lock is released. The default is 30.
--store-endpoints string #: Specifies the etcd address in the format: http[s]://address[:port](,http[s]://address[:port])*. The default is http://127.0.0.1:2379.
--store-ca-file string #: Verify the certificate of the HTTPS-enabled etcd store server using this CA bundle.
--store-cert-file string #: Specifies the certificate file for client identification by the etcd store.
--store-key string #: Specifies the private key file for client identification by the etcd store.
--store-timeout duration #: Specifies the timeout for a etcd request. The default is 5 seconds.
--monitor-port number #: Specifies the port for the shardmand http server for metrics and probes. The default is 15432.
--api-port number #: Specifies the port for the shardmand http api server. The default is 15432.
--version #: Show shardman-utils version information.

Environment

SDM_BACKUP_MODE #: An alternative to setting the --backup-mode option.
SDM_BACKUP_PATH #: An alternative to setting the --backup-path option.
SDM_CLUSTER_NAME #: An alternative to setting the --cluster-name option.
SDM_ETCD_PATH #: An alternative to setting the --etcd-path option.
SDM_FILE #: An alternative to setting the --file option for config update.
SDM_LOG_LEVEL #: An alternative to setting the --log-level option.
SDM_NODES #: An alternative to setting the --nodes option for nodes add and nodes rm.
SDM_RETRIES #: An alternative to setting the --retries option.
SDM_SPEC_FILE #: An alternative to setting the --spec-file option for init.
SDM_STORE_ENDPOINTS #: An alternative to setting the --store-endpoints option.
SDM_STORE_CA_FILE #: An alternative to setting the --store-ca-file option.
SDM_STORE_CERT_FILE #: An alternative to setting the --store-cert-file option.
SDM_STORE_KEY #: An alternative to setting the --store-key option.
SDM_STORE_TIMEOUT #: An alternative to setting the --store-timeout option.
SDM_SESSION_TIMEOUT #: An alternative to setting the --session-timeout option.

Usage

Adding Nodes to a Shardman Cluster

To add nodes to a Shardman cluster, run the following command:

shardmanctl [common_options] nodes add -n|--nodes node_names

You must specify the -n (--nodes) option to pass the comma-separated list of nodes to be added. Nodes can be referred by their hostname or IP address. Hostnames must be correctly resolved on all nodes.

If nodes add command fails during execution, use the cleanup --after-node-operation command to fix possible cluster configuration issues.

Performing Cleanup

By default, cleanup operates in the report-only mode, that is, the following command will only show actions to be done during actual cleanup:

    shardmanctl [common_options] cleanup --after-node-operation|--after-rebalance

To perform the actual cleanup, run the following command:

    shardmanctl [common_options] cleanup -p|--processrepgroups --after-node-operation|--after-rebalance

Displaying the Cluster Topology

cluster topology displays the current cluster topology. The default is the table mode. All cluster nodes will be grouped by the replication groups they belong to. For each node, its status will be displayed.

    shardmanctl [common_options] cluster topology -f|--format table|json|text

Checking shardmand Service on Nodes

daemon check not only checks that shardmand service is running on specified nodes, but also assures those services are configured for the same cluster as shardmanctl:

    shardmanctl [common_options] daemon check -n|--nodes node_names

Removing Nodes from a Shardman cluster

To remove nodes from a Shardman cluster, run the following command:

shardmanctl [common_options] nodes rm -n|--nodes node_names

Specify the -n (--nodes) option to pass the comma-separated list of nodes to be removed.Recreates all partitions of sharded tables

Note

Do not use the cleanup command to fix possible cluster configuration issues after a failure of nodes rm. Redo the nodes rm command instead.

To remove all nodes in a cluster and not care about the data, just reinitialize the cluster. If a removed replication group contains local (non-sharded and non-global) tables, the data is silently lost after the replication group removal.

Getting the Status of Cluster Subsystems

To get a report on the health status of Shardman cluster in a table format for metadata and store subsystems sorted by replication group, run the following command:

    shardmanctl [common_options] status --filter=metadata,store --sort=rg

To get the report in JSON format, use -f|--format=json option (omitted above since table format is used by default). Each detected issue is reported as an Unknown, Warning, Error or Fatal error status. The tool can also report an Operational error, which means there was an issue during the cluster health check. When the command encounters a Fatal or Operational error, it stops further diagnostics. For example, an inconsistency in the store metadata does not allow correct cluster operations and must be handled first.

Outputting the List of Unresolved Distributed Transactions

To view the list of distributed transactions that Shardman built-in monitoring tools failed to resolve, run the following command:

    shardmanctl [common_options] status transactions -r|--repgroup replication_group_name

Each output transaction consists of tx_id (transaction ID), coordinator_id, creation_time and description (error or transaction status). To display the list of transactions for a specific replication group, use the -r|--repgroup option (for all replication groups by default). In case there are no such transactions, returns null value in JSON.

Dumping All Keys from the Store to Debug Error Configuration

After facing an error while using Shardman cluster, to fill in an exhaustive report, it is convinient to dump all specifications that could produce such an error with the following command:

    shardmanctl [common_options] store dump -f|--file filename

Some harmless errors may be shown, but they will not interrupt dumping. If you do not specify the filename, dump will be sent to stdout and may pollute your terminal.

Getting the Current stolon Specification

To get the current stolon specification, which is normally a part of cluster key in the store, use the following command:

    shardmanctl [common_options] store get -a|--alias stolonspec -f|--file filename

If the cluster key is corrupted itself, stolon specification will not be shown either. Instead of using the alias, you may also find out the full cluster data key name (by listing all keys with store keys command), use store get to retrieve it and find the stolon part there. Mind that while using the last option, shardman.config_uuid parameter will not be deleted, which may result in a conflict in later use of this data; for manipulation with stolon specification, it is recommended to use shardmanctl store get -a stolonspec command.

Getting the Cluster and Ladle Key Names For the Current Cluster

To get all key names in the store at once, run the following command:

    shardmanctl [common_options] store keys

It can only be shown in JSON format. It will also print alias names for keys that have them (excluding stolonspec and spec, since they are parts of other keys)

Output Current Cluster Meta Lock Information

You can view information about current cluster meta locks that acquired by any command:

    shardmanctl [common_options] store lock -f|--format json

To get the report in JSON format, use -f|--format=json option (omitted above since text format is used by default). In case the lock does not exists returns Lock not found

Setting a New Spec for the Cluster

To set a new spec part of the cluster specification, run the following command:

    shardmanctl [common_options] store set --alias=spec --file=spec.json

Since spec is a part of cluster data key, it cannot be set with --key. If the provided file is not a valid JSON, the new spec part will not be set.

Backing up a Shardman Cluster

Requirements for backing up and restoring a Shardman cluster using the basebackup command are listed in Section 2.6.1.1.

To backup a Shardman cluster, you can run the following command:

    shardmanctl [common_options] backup --datadir directory [--use-ssh]

You must pass the directory to write the output to through the --datadir option. You can limit the number of running concurrent tasks (pg_receivewal or pg_basebackup commands) by passing the limit through the --maxtasks option.

If --use-ssh is specified shardmanctl recover command will use scp command to restore data. It allows to use backup repository on the local host.

Registering a Shardman Cluster

To register a Shardman cluster in the etcd store, run the following command:

    shardmanctl [common_options] init [-y|--yes] [-f|--spec-file spec_file_name]|spec_text

You must provide the string with the cluster specification. You can do it as follows:

On the command line — do not specify the -f option and pass the string in spec_text.
On the standard input — specify the -f option and pass - in spec_file_name.
In a file — specify the -f option and pass the filename in spec_file_name.

Restoring a Shardman Cluster

shardmanctl can perform either full restore, metadata-only or schema-only restore of a Shardman cluster from a backup created by the backup command.

To perform full restore, you can run the following command:

    shardmanctl [common_options] recover --info file

Pass the file to load information about the backup from through the --info option. In most cases, set this option to point to the backup_info file in the backup directory or to its modified copy.

If you encounter issues with an etcd instance, it makes sense to perform metadata-only restore. To do this, you can run the following command:

   shardmanctl [common_options] recover --dumpfile file --metadata-only

You must pass the file to load the etcd metadata dump from through the --dumpfile option.

If you need to restore only schema information, like: tables, roles and etc. you should specify --schema-only option.

For all kinds of restore, you can specify --timeout for the tool to exit with error after waiting until the cluster is ready or the recovery is complete for the specified number of seconds.

You can specify --shard parameter for restoring only on the single shard.

Before running the recover command, specify DataRestoreCommand and RestoreCommand in the backup_info file. DataRestoreCommand fetches the base backup and restores it to the stolon data directory. RestoreCommand fetches the WAL file and saves it to stolon pg_wal directory. These commands can use the following substitutions:

%p #: Destination path on the server.
%s #: SystemId of the restored database (the same in the backup and in restored cluster).
%f #: Name of the WAL file to restore.

stolon keeper thread runs both commands on each node in the cluster. Therefore:

Make the backup accessible to these nodes (for example, by storing it in a shared filesystem or by using a remote copy protocol, such as SFTP).
Commands to fetch the backup are executed as the operating system user under which stolon daemons work (usually postgres), so set the permissions for the backup files appropriately.

These examples show how to specify RestoreCommand and DataRestoreCommand:

If a backup is available through a passwordless SCP, you can use:

 "DataRestoreCommand": "scp -r user@host:/var/backup/shardman/%s/backup/* %p",
 "RestoreCommand": "scp user@host:/var/backup/shardman/%s/wal/%f %p"

If a backup is stored on NFS and available through /var/backup/shardman path, you can use:

 "DataRestoreCommand": "cp -r /var/backup/shardman/%s/backup/* %p",
 "RestoreCommand": "cp /var/backup/shardman/%s/wal/%f %p"

Backing up a Shardman Cluster Using `probackup` Command

Requirements for backing up and restoring a Shardman cluster using the probackup command are listed in Section 2.6.3.1.

For example, following these requirements, on the backup host:

  groupadd postgres
  useradd -m -N -g postgres -r -d /var/lib/postgresql -s /bin/bash

Then add SSH keys to provide passwordless SSH connection between the backup host and Shardman cluster hosts. Then on the backup host:

  apt-get install pg-probackup shardman-utils
  mkdir -p directory
  chown postgres:postgres directory -R
  shardmanctl [common_options] probackup init --backup-path=directory --etcd-path=directory/etcd --remote-user=postgres --remote-port=22
  shardmanctl [common_options] probackup archive-command --backup-path=directory --remote-user=postgres --remote-port=22

If all the requirements are met, then run the backup subcommand for the cluster backup:

shardmanctl [common_options] probackup backup --backup-path=directory --etcd-path=directory --backup-mode=MODE

You must pass the directories through the --backup-path and --etcd-path options and backup mode through --backup-mode. Full and delta backups are available with FULL, DELTA, PTRACK and PAGE values. Also it is possible to specify backup compression options through --compress, --compress-algorithm and --compress-level flags, as well as specify --remote-port and --remote-user flags. You can limit the number of running concurrent tasks when doing backup by passing the limit through the --maxtasks flag.

By default, copying data via SSH is used to create a backup. To copy data to a mounted partition instead, use the --storage-type option with the mount value. This value will be automatically used in the restore process.

You can also copy data to an S3-compatible object storage. To do this, use the --storage-type option with the S3 value. When this value is used, it is required to specify the directory for pg_probackup logs. You can do it either by specifying --log-directory for each command or set the environment variable SDM_LOG_DIRECTORY, for example:

export SDM_LOG_DIRECTORY=/backup/logs

If you are going to perform backup/restore only for an S3-compatible object storage, you can also set an environment variable instead of specifying --storage-type in each probackup command:

export SDM_STORAGE_TYPE=S3

Restoring a Shardman Cluster using probackup command

shardmanctl in probackup mode can perform either full restore, metadata-only or schema-only restore of a Shardman cluster from a backup created by the probackup backup command.

To perform full or partial restore, firstly you must select needed backup to restore from. To show list of available backups run the following command:

shardmanctl [common_options] probackup show --backup-dir=path --format=format [--archive ] [-i|--backup-id backup-id] [--instance instance]

The output should be a list of backups with their IDs in a table or JSON format. Then pick the needed backup ID and run the probackup restore command.

shardmanctl [common_options] probackup restore --backup-dir=path --backup-id=id

Pass the path to the repo through the --backup-dir option and backup ID througt --backup-id flag.

If you encounter issues with an etcd instance, it makes sense to perform metadata-only restore. To do this, you can run the following command:

shardmanctl [common_options] probackup restore --backup-dir=path --backup-id=id --metadata-only

If you need to restore only schema information, like: tables, roles and etc. you should specify --schema-only option.

For both kinds of restore, you can specify --timeout for the tool to exit with error after waiting until the cluster is ready or the recovery is complete for the specified number of seconds.

You can specify --shard parameter for restoring only on the single shard.

Also you can specify --recovery-target-time option for Point-in-Time Recovery. In this case Shardman finds the closest syncpoint to specified timestamp and suggests restoring on the found LSN. You can also specify --wal-limit to limit the number of WAL segments to be processed.

Important

Before restoring a Shardman cluster, make sure that the cluster is up by executing the shardmanctl status command. If the output shows errors, performing the restore can result in the cluster becoming unavailable. First, fix the errors by reinitializing the cluster and restoring the etcd metadata. Then you can proceed to restoring the cluster from backup.

Reinitializing Replicas

If replicas are in an incorrect state, you can reset them using the shardmanctl command:

shardmanctl [common_options] shard --shard=shard_name replicas reinit

This command determines the nodes on which replicas of the specified shard are running and sends a request to shardmand on these nodes. After receiving this request, shardmand clears the postgres data directory and restarts the keeper thread that is responsible for managing the replica. After that, the replicas are restarted and begin to receive data from the corresponding primary.

Examples

Initializing the Cluster

To initialize a Shardman cluster that has the cluster0 name, uses an etcd cluster consisting of n1,n2 and n3 nodes listening on port 2379, ensure proper settings in the spec file sdmspec.json and run:

$ shardmanctl --store-endpoints http://n1:2379,http://n2:2379,http://n3:2379 init -f sdmspec.json

Getting the Cluster Connection String

To get the connection string for a Shardman cluster that has the cluster0 name, uses an etcd cluster consisting of n1,n2 and n3 nodes listening on port 2379, run:

 $ shardmanctl --store-endpoints http://n1:2379,http://n2:2379,http://n3:2379 getconnstr
 
 dbname=postgres host=n1,n4,n2,n1,n1,n2,n4,n3 password=yourpasswordhere port=5432,5433,5432,5433,5432,5433,5432,5433 user=postgres

To add replicas to getconnstr, use --all.

Getting the Cluster Status

Here is a sample status output from shardmanctl with OK and Error statuses:

        $ shardmanctl status --filter store,shardmand,rg --sort=node
        
┌──────────────────────────────────────────────────────────────────────────────────────────────────────────┐
│                                            == STORE STATUS ==                                            │
├────────────┬─────────────────────────────────────────────┬───────────────────────┬───────────────────────┤
│   STATUS   │                   MESSAGE                   │   REPLICATION GROUP   │          NODE         │
├────────────┼─────────────────────────────────────────────┼───────────────────────┼───────────────────────┤
│     OK     │ etcd store is OK                            │                       │                       │
└────────────┴─────────────────────────────────────────────┴───────────────────────┴───────────────────────┘
┌──────────────────────────────────────────────────────────────────────────────────────────────────────────┐
│                                          == SHARDMAND STATUS ==                                           │
├────────────┬─────────────────────────────────────────────┬───────────────────────┬───────────────────────┤
│   STATUS   │                   MESSAGE                   │   REPLICATION GROUP   │          NODE         │
├────────────┼─────────────────────────────────────────────┼───────────────────────┼───────────────────────┤
│     OK     │ shardmand on node 56d819b4e9e4 is OK        │                       │      56d819b4e9e4     │
├────────────┼─────────────────────────────────────────────┼───────────────────────┼───────────────────────┤
│     OK     │ shardmand on node 6d0aabd50acc is OK        │                       │      6d0aabd50acc     │
└────────────┴─────────────────────────────────────────────┴───────────────────────┴───────────────────────┘
┌───────────────────────────────────────────────────────────────────────────────────────────────────────────┐
│                                       == REPLICATION GROUP STATUS ==                                      │
├────────────┬──────────────────────────────────────────────┬───────────────────────┬───────────────────────┤
│   STATUS   │                    MESSAGE                   │   REPLICATION GROUP   │          NODE         │
├────────────┼──────────────────────────────────────────────┼───────────────────────┼───────────────────────┤
│     OK     │ Replication group clover-1-56d819b4e9e4 is   │ clover-1-56d819b4e9e4 │                       │
│            │ OK                                           │                       │                       │
├────────────┼──────────────────────────────────────────────┼───────────────────────┼───────────────────────┤
│            │ Replication connection is down for slave     │                       │                       │
│    Error   │ 6d0aabd50acc:5442 in replication group       │ clover-1-6d0aabd50acc │   6d0aabd50acc:5442   │
│            │ clover-1-6d0aabd50acc                        │                       │                       │
└────────────┴──────────────────────────────────────────────┴───────────────────────┴───────────────────────┘

Rewriting stolon Specification

First, get the list of available keys in the store using the following command:

      $ shardmanctl store keys
        
{
 "Key": "shardman/cluster0/data/cluster",
 "Alias": "cluster"
}{
 "Key": "shardman/cluster0/data/shardmand/56d819b4e9e4"
}{
...
 "Key": "shardman/cluster0/stolon/remoteLogs/6d0aabd50acc/clover-1-6d0aabd50acc/keeper_1/error"
}

Get stolon configuration from the store and save it in the stolonspec.json file with the command

      $ shardmanctl store get -a stolonspec -f stolonspec.json

Apply the necessary changes to the file and upload the new specification using shardmanctl config update. Mind that shardman.config_uuid parameter is deleted with shardmanctl store get -a stolonspec and not with shardmanctl store get -k full/path/to/clusterspec; using spec with existing shardman.config_uuid will result in a conflict.

Important

Do not use store set command to update cluster configurations because it does not apply a new specification on all nodes, it only writes it to the store. For the above example with stolon specification, shardmanctl config update is acceptable.

To double-check, you can get the cluster key with new StolonSpec by the full key name (which was shown earlier with store keys command):

      $ shardmanctl store get -k shardman/cluster0/data/cluster
        
{
  "FormatVersion": 1,
  "Spec": {
    "PgSuAuthMethod": "md5",
    "PgSuPassword": "12345",
    "PgSuUsername": "postgres",
    "PgReplAuthMethod": "md5",
    "PgReplPassword": "12345",
    "PgReplUsername": "repluser",
    "ShardSpec": {
      ...
}

Adding Nodes to the Cluster

To add n1,n2, n3 and n4 nodes to the cluster, run:

$ shardmanctl --store-endpoints http://n1:2379,http://n2:2379,http://n3:2379 nodes add -n n1,n2,n3,n4

Important

The number of nodes being added must be a multiple of Repfactor + 1 if cross placement policy is used.

Removing Nodes from the Cluster

To remove n1 and n2 nodes from the cluster0 cluster, run:

      $ shardmanctl --store-endpoints http://n1:2379,http://n2:2379,http://n3:2379 nodes rm -n n1,n2

If cross placement policy is used, then the clovers that contain them will be deleted along with the nodes.

Executing a Query on All Replication Groups

To execute the select version() query on all replication groups, run:

 $ shardmanctl --store-endpoints http://n1:2379,http://n2:2379,http://n3:2379 forall --sql 'select version()'
 
 Node 1 says:
 [PostgreSQL 13.1 on x86_64-pc-linux-gnu, compiled by gcc (Ubuntu 9.3.0-17ubuntu1~20.04) 9.3.0, 64-bit]
 Node 4 says:
 [PostgreSQL 13.1 on x86_64-pc-linux-gnu, compiled by gcc (Ubuntu 9.3.0-17ubuntu1~20.04) 9.3.0, 64-bit]
 Node 3 says:
 [PostgreSQL 13.1 on x86_64-pc-linux-gnu, compiled by gcc (Ubuntu 9.3.0-17ubuntu1~20.04) 9.3.0, 64-bit]
 Node 2 says:
 [PostgreSQL 13.1 on x86_64-pc-linux-gnu, compiled by gcc (Ubuntu 9.3.0-17ubuntu1~20.04) 9.3.0, 64-bit]

Performing Rebalance

To rebalance sharded tables in the cluster0 cluster, run:

 $ shardmanctl --store-endpoints http://n1:2379,http://n2:2379,http://n3:2379 rebalance

Updating PostgreSQL Configuration Settings

To set the max_connections parameter to 200 in the cluster, create the spec file (for instance, ~/stolon.json) with the following contents:

 {
   "pgParameters": {
     "max_connections": "200"
   }
 }

Then run:

 $ shardmanctl --store-endpoints http://n1:2379,http://n2:2379,http://n3:2379 config update -p -f ~/stolon.json

Since changing max_connections requires a restart, DBMS instances are restarted by this command.

Performing Backup and Recovery

To create a backup of the cluster0 cluster using etcd at etcdserver listening on port 2379 and store it in the local directory /var/backup/shardman, run:

$ shardmanctl --store-endpoints http://etcdserver:2379 backup --datadir=/var/backup/shardman --use-ssh

Assume that you are performing a recovery from a backup to the cluster0 cluster using etcd at etcdserver listening on port 2379 and you take the backup description from the /var/backup/shardman/backup_info file. Edit the /var/backup/shardman/backup_info file, set DataRestoreCommand , RestoreCommand as necessary and run:

$ shardmanctl --store-endpoints http://etcdserver:2379 recover --info /var/backup/shardman/backup_info

For metadata-only restore, run:

$ shardmanctl --store-endpoints http://etcdserver:2379 recover --metadata-only --dumpfile /var/backup/shardman/etcd_dump

For schema-only restore, run:

$ shardmanctl --store-endpoints http://etcdserver:2379 recover --schema-only --dumpfile /var/backup/shardman/etcd_dump

For single shard restore, run:

$ shardmanctl --store-endpoints http://etcdserver:2379 recover  --info /var/backup/shardman/backup_info --shard shard_1

Performing Backup and Recovery with probackup Command

To create a backup of the cluster0 cluster using etcd at etcdserver listening on port 2379 and store it in the local directory /var/backup/shardman, first initialize the backups repository with the init subcommand:

$ shardmanctl --store-endpoints http://etcdserver:2379 probackup init --backup-path=/var/backup/shardman --etcd-path=/var/backup/etcd_dump

Then add and enable archive_command with the archive-command subcommand:

$ shardmanctl --store-endpoints http://etcdserver:2379 probackup archive-command add --backup-path=/var/backup/shardman

If the repository is successfully initialized and archive-command successfully added, create a FULL backup with the backup subcommand:

$ shardmanctl --store-endpoints http://etcdserver:2379 probackup backup --backup-path=/var/backup/shardman --etcd-path=/var/backup/etcd_dump --backup-mode=FULL --compress --compress-algorithm=zlib --compress-level=5

To create DELTA, PTRACK or PAGE backup, run the backup subcommand with DELTA, PTRACK or PAGE value of the --backup-mode option:

$ shardmanctl --store-endpoints http://etcdserver:2379 probackup backup --backup-path=/var/backup/shardman --etcd-path=/var/backup/etcd_dump --backup-mode=DELTA --compress --compress-algorithm=zlib --compress-level=5

To show the created backup ID, run show subcommand:

$ shardmanctl --store-endpoints http://etcdserver:2379 probackup show --backup-path=/var/backup/shardman --format=table

┌────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────┐
│                                                                                                      == BACKUP ID 'S88FRO'                                                                                                     │
│                                                                                                               ==                                                                                                               │
├───────────────────────────────────┬───────────────────────────────────┬───────────────────────────────────┬────────────┬────────────┬────────────┬────────────┬────────────┬────────────┬────────────┬────────────┬────────────┤
│              INSTANCE             │                HOST               │           RECOVERY TIME           │    MODE    │  WAL MODE  │     TLI    │    DATA    │     WAL    │   Z-RATIO  │  START LSN │  STOP LSN  │   STATUS   │
├───────────────────────────────────┼───────────────────────────────────┼───────────────────────────────────┼────────────┼────────────┼────────────┼────────────┼────────────┼────────────┼────────────┼────────────┼────────────┤
│              shard-1              │               n1                  │       2024-02-02 14:19:05+00      │    FULL    │   ARCHIVE  │     1/0    │  42.37MiB  │    16MiB   │    1.00    │  0/C000028 │  0/D0018B0 │     OK     │
├───────────────────────────────────┼───────────────────────────────────┼───────────────────────────────────┼────────────┼────────────┼────────────┼────────────┼────────────┼────────────┼────────────┼────────────┼────────────┤
│              shard-2              │               n2                  │       2024-02-02 14:19:05+00      │    FULL    │   ARCHIVE  │     1/0    │  42.38MiB  │    16MiB   │    1.00    │  0/C000028 │  0/D001E00 │     OK     │
└───────────────────────────────────┴───────────────────────────────────┴───────────────────────────────────┴────────────┴────────────┴────────────┴────────────┴────────────┴────────────┴────────────┴────────────┴────────────┘

In PTRACK backup mode, Shardman tracks page changes on the fly. Continuous archiving is not necessary for it to operate. Each time a relation page is updated, this page is marked in a special PTRACK bitmap. Tracking implies some minor overhead on the database server operation, but speeds up incremental backups significantly.

If you are going to use PTRACK backups, complete the following additional steps:

Preload the ptrack shared library on each node. This can be done by adding the ptrack value to the shared_preload_libraries parameter.

Сreate the PTRACK extension on each cluster node:

                            $ shardmanctl --store-endpoints http://etcdserver:2379 forall --sql "create extension ptrack"

To enable tracking page updates, set the ptrack.map_size parameter as follows:
```
                            $ shardmanctl --store-endpoints http://etcdserver:2379 update '{"pgParameters":{"ptrack.map_size":"64"}}'
```
For optimal performance, it is recommended to set ptrack.map_size to N/1024, where N is the maximum size of the cluster node, in MB. If you set this parameter to a lower value, PTRACK is more likely to map several blocks together, which leads to false-positive results when tracking changed blocks and increases the incremental backup size as unchanged blocks can also be copied into the incremental backup. Setting ptrack.map_size to a higher value does not affect PTRACK operation, but it is not recommended to set this parameter to a value higher than 1024.

To validate the created backup, run validate subcommand:

$ shardmanctl --store-endpoints http://etcdserver:2379 probackup validate --backup-path=/var/backup/shardman --backup-id=RFP1FI

Assume that you are performing a recovery from a backup to the cluster0 cluster using etcd at etcdserver listening on port 2379 and you take the backup ID from the show command:

$ shardmanctl --store-endpoints http://etcdserver:2379 probackup restore --backup-path=/var/backup/shardman --backup-id=RFP1FI

Finally we need to enable archive_command back.

$ shardmanctl --store-endpoints http://etcdserver:2379 probackup archive-command add --backup-path=/var/backup/shardman

For metadata-only restore, run:

$ shardmanctl --store-endpoints http://etcdserver:2379 probackup restore --metadata-only --backup-path=/var/backup/shardman --backup-id=RFP1FI

For metadata-only restore, run:

$ shardmanctl --store-endpoints http://etcdserver:2379 probackup restore --schema-only --backup-path=/var/backup/shardman --backup-id=RFP1FI

For single shard restore, run:

$ shardmanctl --store-endpoints http://etcdserver:2379 probackup restore --backup-path=/var/backup/shardman --backup-id=RFP1FI --shard shard_1

For Point-in-Time Recovery, run:

$ shardmanctl --store-endpoints http://etcdserver:2379 probackup restore --metadata-only --backup-path=/var/backup/shardman --backup-id=RFP1FI --recovery-target-time='2006-01-02 15:04:05' -s

Loading Data from a Text File

To load data into a Shardman cluster, run the following command:

$ shardmanctl --store-endpoints http://etcdserver:2379 load --file=/var/load/data.tsv --table=mytable --source file --format text -j 8

In this example, data is loaded from the /var/load/data.tsv data file (tab-delimited) into the table mytable in 8 parallel threads. You can use schema.table as the table name.

Loading data from PostgreSQL table

To load data into a Shardman cluster from a PostgreSQL table, run the following command:

$ shardmanctl --store-endpoints http://etcdserver:2379 load -t desttable --source postgres --source-connstr "dbname=db host=srchost port=srcport user=login password=passwd" --source-table sourcetable -j 8

In this example, data is loaded from the table sourcetable into the desttable table in 8 parallel threads. You can use schema.table as table names.

Loading Data with a Schema from PostgreSQL

To load data with a schema into Shardman cluster from PostgreSQL, run the following command:

$ shardmanctl --store-endpoints http://etcdserver:2379 load --schema load_schema.yaml

The file load_schema.yaml has the folowing format:

version: "1.0"
migrate:
  connstr: "dbname=workdb host=workhost port=workport user=workuser password=workpassword"
  jobs: 8
  batch: 1000
  options:
    - create_schema
    - create_table
    - create_index
    - create_sequence
    - create_foreign_key
    - create_role
    - copy_ownership
    - copy_grants
    - truncate_table
    - skip_no_pkey_tables
    - skip_create_index_error
    - skip_create_extension_error
    - skip_load_errors
    - skip_create_foreign_key_error
    - skip_create_role_error
    - skip_copy_grants_error
    - skip_copy_ownership_error
  schemas:
    - name: public
      all: false
      tables:
        - name: tab1
          type: sharded
          partitions: 6
          distributedby: id
          priority: 3
        - name: tab2
          type: global
        - name: tab3
          type: sharded
          partitions: 6
          distributedby: field_id
          colocatewith: tab1
        - name: table4
            type: global
            source: schema.view
            source_pk: field_id
        - name: table5
            type: global
            source: schema.func(arg)
            source_pk: field_id
    - name: schema2
      all: false
      default_type: sharded
      default_partitions: 6
      tables:
        - name: table1
          distributedby: field_id
          priority: 2
        - name: table2
          type: global
        - name: table3
          source: schema.view
          distributedby: field_id
          priority: 3
        - name: table4
          distributedby: field_id
          source: schema.func(arg)
        - name: table5
          source: schema."complex.""table.name"
          distributedby: field_id
    - name: schema3
      all: true
      skip_tables: [table1, table2, table3]
  roles:
    - name: test_user1
      password: test_password
    - name: test_user2

The migrate.jobs value defines the number of parallel data loader processes.

The migrate.batch value is the number of rows in one batch (recommended value is 1000).

The migrate.schemas section defines an array of source database schemas that you are working with. All other schemas will be skipped.

If the all value is set to true, then all tables from the current schema will be migrated (with global type by default). If a table is listed in the migrate.schemas.tables array, then the target table type must be explicitly specified for it. Two types of tables are currently supported: global and sharded. Global tables are loaded first, then sharded tables and at the end sharded tables with the colocatedwith parameter. The order of loading tables of the same type can be changed using priority option.

The migrate.schemas.skip_tables section defines an array of table names that will be skipped when the schema is loaded even if the all parameter is set to true.

For sharded tables, the following attributes must be set: distributedby (specifies the name of the column to use for the table partitioning) and partitions (number of partitions that will be created for this table). Optionally, for sharded tables colocatewith attribute can be set (name of the table to colocate with). Shardman will try to place partitions of the created table with the same partition key on the same nodes as the corresponding partitions of the table specified by colocatewith.

You can specify the table default_type option for a schema: global or sharded (default: global). For the sharded type you can also specify the default_partitions option (default: 20). If you set default_type to sharded, you need to specify the distributedby option for each table.

The source option for a table should include the schema and table source: schema.source. The source can be a table, view or function. For example: public.table, public.view, public.func(arg). If you set the source view or function for a global table, you should specify source_pk to set the primary key for this table. If source is not specified or contains the name of a table, you can also specify source_pk to create a primary key or override the existing one.

The priority option for table determines the order in which the tables of the same type are loaded. Tables with higher priority are loaded earlier. Default priority value is 0.

The migrate.roles section defines an array of role names and passwords that will be copied from the source database if create_role is specified.

The schema supports the following options:

create_schema — create database schemas if they do not exist.
create_table — create tables if they do not exist.
create_index — create indexes after creating tables.
create_sequence — create sequences if they do not exist.
create_foreign_key — create foreign keys after creating tables.
truncate_table — truncate tables before data load.
create_role — create global roles defined in migrate.roles and copy role parameters from the source database.
copy_grants — copy access privileges from the source database.
copy_ownership — change of table owners to the owner in the source database.
skip_no_pkey_tables — skip tables without primary keys.
skip_create_index_error — skip index creation errors.
skip_create_extension_error — skip extension creation errors.
skip_load_errors — continue loading if errors occur.
skip_create_foreign_key_error — skip foreign key creation errors.
skip_create_role_error — skip role creation errors.
skip_copy_ownership_error — skip table owner changing errors.
skip_copy_grants_error — skip errors when copying access privuleges from the source database.

Initialization and Running Benchmarks

To initialize a benchmark via shardmanctl using pgbench with the shardman schema, scale=1000, partitions=40, run:

$ shardmanctl bench init --schema-type=shardman --scale=1000 --partitions=40

To run an initialized benchmark for the same shardman schema, number of jobs=4, number of clients=10, duration in seconds=60 and full pgbench output, use:

$ shardmanctl bench run --schema-type=shardman --jobs=4 --client=10 --time=60 --full-output

To initialize a benchmark with the custom schema from file schema.psql with scale=1000 run:

$ shardmanctl bench init --schema-type=custom --schema-file=schema.psql --scale=1000

To run an initialized benchmark with the custom schema and custom transaction script from script.psql with the number of jobs=4, number of clients=10, duration in seconds=60, use:

$ shardmanctl bench run --schema-type=custom --file=script.psql --jobs=4 --client=10 --time=60

To clean up a PostgreSQL database of tpc-b tables, use:

$ shardmanctl bench cleanup

Benchmark Generation Scripts

To generate a benchmark sequence via shardmanctl from the config file=cfg.yaml and output the result to file=script.sh, run:

$ shardmanctl bench generate --config=cfg.yaml --output-file=script.sh

Configuration file example:

    benches:
    - schema_type: single
      init_flags: "-s 3"
      run_flags: "-n -P 10 -c 10 -j 4 -T 10"
    - schema_type: simple
      init_flags: "-s 4"
      run_flags: "-n -P 10 -c 20 -j 4 -T 10"
      partitions: 100
    - schema_type: shardman
      init_flags: "-s 5"
      run_flags: "-n -P 10 -c 20 -j 4 -T 10"
    - schema_type: custom
      init_flags: "-s 6"
      schema_file: "schema.psql"