Work with resource groups

Daria Barysheva

The Monitoring → Resource groups page in the ADB Control web interface displays the resource groups that are used in ADB clusters connected to the monitoring system. This page includes two tabs Statistic and Configuration, each of which is described in detail below.

Statistic

The Monitoring → Resource groups → Statistic tab displays the statistics of the resource group usage in the selected ADB cluster. The Last updated field contains date and time of the last data update on the page in the DD/MM/YYYY HH:mm:ss format.

adbc rg stat dark
The Monitoring → Resource groups → Statistic tab
adbc rg stat light
The Monitoring → Resource groups → Statistic tab

On the tab, there is a table with the following information on resource groups.

The "Monitoring → Resource groups → Statistic" tab fields
Field Description

Group ID

A unique resource group identifier

Resource group

A resource group name

Running

A number of transactions currently running within the current resource group

Queueing

A number of transactions currently queued within the current resource group

Queued

A total number of transactions queued within the current resource group over the entire observation period

Executed

A total number of transactions executed within the current resource group over the entire observation period

Total queued duration

The total time that all transactions were queued within the current resource group over the entire observation period (in hours, minutes, seconds)

Above the table with a list of resource groups, the Cluster filter is located. You can use this filter to select the ADB cluster for which you want to display data in the table. The default cluster for all such filters is defined in the ADB Control settings.

Configuration

List of resource groups

The Monitoring → Resource groups → Configuration tab displays the current configurations of existing resource groups.

adbc rg config dark
The Monitoring → Resource groups → Configuration tab
adbc rg config light
The Monitoring → Resource groups → Configuration tab

On the tab, there is a table with the following information on resource groups.

The "Monitoring → Resource groups → Configuration" tab fields
Field Description

Group ID

A unique resource group identifier

Resource group

A resource group name

Concurrency

The maximum number of concurrent transactions that are permitted within the resource group. If the limit is reached, new commands will be queued

CPU ratio limit

The percentage of the reserved CPU resources of the entire cluster (specified in the gp_resource_group_cpu_limit parameter) that is available to the resource group. The sum of this parameter values for all resource groups should be <= 100. The parameter is not compatible with CPU set — fill only one of two parameters

CPU set

The CPU cores that are reserved for the resource group on the master and segment hosts. If the parameter is set, all SQL commands within the resource group will only use the specified cores. If the resource group does not use cores, they will be idle. The parameter is not compatible with CPU ratio limit — fill only one of two parameters. The -1 value means that CPU ratio limit is used. Example: 1,3-4

Memory limit

The percentage of the reserved RAM resources of the entire cluster (specified in the gp_resource_group_memory_limit parameter) that is available to the resource group. The sum of this parameter values for all resource groups should be <= 100

Memory shared quota

The percentage of the reserved RAM resources of the resource group to be shared between all transactions running within that group. The amount of memory equal to (100-<Memory shared quota>)/<Concurrency> will be guaranteed to be available for each transaction in the group

Memory spilled ratio

The memory usage threshold for memory-intensive transactions. When a transaction reaches the specified threshold, spill files are created

Auditor

The memory auditor that is used for the resource group. The default value is vmtracker

Above the table with a list of resource groups, the Cluster filter is located. You can use this filter to select the ADB cluster for which you want to display data in the table. The default cluster for all such filters is defined in the ADB Control settings.

NOTE

For more information on resource group configuration parameters, see Greenplum documentation.

Edit a resource group configuration

To edit the resource group configuration, follow the steps:

  1. In the ADB cluster for which you plan to modify the resource group settings, assign the super-user rights to the role that is used in the JDBC connection to that cluster. The user is defined on the Configuration → Clusters page (by default adcc). To set super-user rights, run the following query:

    ALTER ROLE adcc SUPERUSER;

    Skipping the first step will cause the following error in step 4.

    adbc rg error dark
    The error that occurs for non super-users
    adbc rg error light
    The error that occurs for non super-users

  2. Click the edit dark edit light icon in the Action column on the Monitoring → Resource groups → Configuration tab.

    adbc rg edit1 dark
    Switch to editing a resource group configuration
    adbc rg edit1 light
    Switch to editing a resource group configuration

  3. In the window that opens, edit necessary data. Descriptions of all fields are availble above. A controller name in the Memory auditor field value cannot be changed.

    adbc rg edit2 dark
    Edit the resource group
    adbc rg edit2 light
    Edit the resource group

  4. Click Apply. As a result, the resource group data is updated on the Monitoring → Resource groups → Configuration tab.

    adbc rg edit3 dark
    The resource group data is updated
    adbc rg edit3 light
    The resource group data is updated
IMPORTANT

  • The resource group update is available only to ADB Control users with appropriate permissions (see Change resource group configuration in the Authorization article).

  • The resource group update is recorded as a part of the operation audit under the operation name Update (for the Resource group object type). You can view audit events on the Audit → Operations tab.

Resource group details

To view the resource group details, click the group identifier (Group ID) in the table on the Monitoring → Resource groups → Statistic or Monitoring → Resource groups → Configuration tab.

adbc rg stat2 dark
Select a resource group
adbc rg stat2 light
Select a resource group

The following page contains multiple sections that are described below.

At the top of the page, the resource group identifier and name are displayed.

adbc rg details header dark
Header of the page with resource group details
adbc rg details header light
Header of the page with resource group details

Overview

The Overview section displays the general resource group information. All fields are described in detail above.

adbc rg details overview1 dark
The Overview section
adbc rg details overview1 light
The Overview section

Commands

The Commands section lists the commands that were started within the current resource group and currently have the following statuses:

  • Queued — the commands on which information about the actual start of execution has not yet received.

  • Running — the commands that are currently running.

  • Cancelling — the commands that are currently cancelling (as a part of the transaction cancellation or termination).

The "Commands" tab fields
Field Description

Command ID

A unique command identifier, which includes:

  • cluster start timestamp;

  • session number;

  • command number within the current session;

  • hash code based on the cluster name (can be negative).

To open the page with command details, click the command identifier

SQL ID

A common identifier for SQL commands with the same structure

Command text

The first symbols of a command text. To view the full text (in case of long queries), hover the mouse over the field value

Status

A command status. Possible values are listed above

Planner

A name of the query optimizer that is used to produce the query execution plan. Possible values:

  • GPORCA — GPORCA is used (optimizer = on).

  • Fallbacked — after the attempt to produce a plan via GPORCA (optimizer = on) failed, Postgres query optimizer is used.

  • Legacy — Postgres query optimizer is used (optimizer = off).

Submitted

The timestamp when the user submitted the command (in the DD/MM/YYYY HH:mm:ss format)
adbc rg details overview2 dark
The Commands section
adbc rg details overview2 light
The Commands section

In the column headers of the table with a list of commands, there are filters that you can use to select specific data. To open a filter, click the openside dark openside light icon. For those columns where the set of possible values is limited (e.g. Status), you can select a value from the drop-down list. For some columns (e.g. Command ID), the search value should be entered.

The filtered dark filtered light icon means that a filter is defined for the column. To reset all filters, click Reset.

Usage

The Usage section displays the statistics of system resource consumption by the selected resource group. This section includes two tabs:

  • Hosts

  • Segments

The Hosts tab displays data for each cluster host.

The "Usage → Hosts" tab fields
Field Description

Hostname

A host name

CPU

The percentage of CPU resource usage by the resource group on the current host

Memory quota

The ratio of used and available amount of RAM that is guaranteed fixed for resource group transactions on the current host

Memory shared

The ratio of used and available amount of RAM that is commonly used by resource group transactions on the current host

Memory

The ratio of total used and available amount of RAM on the current host
adbc rg details hosts dark
The Usage → Hosts tab
adbc rg details hosts light
The Usage → Hosts tab

The Segments tab displays data for each cluster segment. In comparison with the Hosts tab, the table contain one additional field Segment ID — a uniqie segment identifier (gp_segment_configuration.content). Other parameters are identical with the only difference that the data is displayed for each segment separately.

adbc rg details seg dark
The Usage → Segments tab
adbc rg details seg light
The Usage → Segments tab
Found a mistake? Seleсt text and press Ctrl+Enter to report it