Command Center Agent Parameters

Post date: Nov 08, 2012 8:19:57 PM

The $MASTER_DATA_DIRECTORY/gpperfmon/conf/gpperfmon.conf stores configuration parameters for the Command Center agents. For configuration changes to these options to take effect, you must save gpperfmon.conf and then restart Greenplum Database server (gpstop -r). 

To enable the Command Center agents within Greenplum Database server, you must also set the “Greenplum Database Server Configuration Parameters” .

The following parameters must be uncommented and set in the server configuration file (postgresql.conf) in order to enable the Command Center data collection agents. 

gp_enable_gpperfmon and gpperfmon_port must be set in both the master and segment postgresql.conf files. gp_enable_gpperfmon and gp_enable_gpperfmon only need to be set in the master postgresql.conf file.

After changing these settings, the Greenplum Database instance must be restarted for the changes to take effect.


Turns on the Command Center data collection agent for a segment. Must be set in all postgresql.conf files (master and all segments).


The default port for the Command Center agents is 8888, but you can set this parameter to a different port if required (master and all segments).


Sets the frequency in seconds that the Greenplum Database server processes send query execution updates to the Command Center agent processes.


This parameter is enabled by default and must remain enabled. It allows the use of external tables that execute OS commands or scripts on the segment hosts. The Command Center agents use this type of external tables to collect current system


Specifies the time in seconds between updates from Command Center agents on all segments. Valid values are 10, 15, 20, 30, and 60. Defaults to 15 seconds.

If you prefer a less granular view of performance, or want to collect and analyze minimal amounts of data for system metrics, choose a higher quantum. To collect data more frequently, choose a lower value.


Specifies the minimum query run time in seconds for statistics collection. Command Center logs all queries that run longer than this value in the queries_history table. For queries with shorter run times, no historical data is collected. Defaults to 20 seconds. If you know that you want to collect data for all queries, you can set this parameter to a low value. Setting the minimum query run time to zero, however, collects data even for the numerous queries run by Command Center itself, creating a large amount of data that may not be useful.


Specifies the minimum iterator run time in seconds for statistics collection. Command Center logs all iterators that run longer than this value in the iterators_history table. For iterators with shorter run times, no data is collected. Minimum value is 10 seconds. 

This parameter’s value must always be equal to, or greater than, the value of min_query_time. Setting min_detailed_query_time higher than min_query_time allows you to log detailed query plan iterator data only for especially complex, long-running queries, while still logging basic query data for shorter queries. Given the complexity and size of iterator data, you may want to adjust this parameter according to the size of data collected. If the iterators_* tables are growing to excessive size without providing useful information, you can raise the value of this parameter to log iterator detail for fewer queries.


Specifies a directory location for Command Center log files. Default is $MASTER_DATA_DIRECTORY/gpperfmon/logs.


This parameter is not included in gpperfmon.conf, but it may be added to this file for use with Greenplum Command Center. To prevent the log files from growing to excessive size, you can add the max_log_size parameter to gpperfmon.conf. The value of this parameter is measured in bytes. For example:

max_log_size = 10485760

With this setting, the log files will grow to 10MB before the system rolls over to a new log file.


This parameter allows you to specify additional host names for the standby master. For example, if the standby master has two NICs, you can enter:

smdw_aliases= smdw-1,smdw-2

This optional fault tolerance parameter is useful if the Greenplum Command Center loses connectivity with the standby master. Instead of continuously retrying to connect to host smdw, it will try to connect to the NIC-based aliases of smdw-1 and/or smdw-2. This ensures that the Command Center Console can continuously poll and monitor the standby master.