Monitoring Greenplum Database Log Files

A newer version of this documentation is available. Use the version menu above to view the most up-to-date release of the Greenplum 5.x documentation.

Monitoring Greenplum Database Log Files

Know the location and content of system log files and monitor them on a regular basis and not just when problems arise.

The following table shows the locations of the various Greenplum Database log files. In file paths, date is a date in the format YYYYMMDD, instance is the current instance name, and n is the segment number.

Path Description
/var/gpadmin/gpadminlogs/* Many different types of log files, directory on each server
/var/gpadmin/gpadminlogs/gpstart_date.log start log
/var/gpadmin/gpadminlogs/gpstop_date.log stop log
/var/gpadmin/gpadminlogs/gpsegstart.py_idb*gpadmin_date.log segment start log
/var/gpadmin/gpadminlogs/gpsegstop.py_idb*gpadmin_date.log segment stop log
/var/gpdb/instance/datamaster/gpseg-1/pg_log/startup.log instance start log
/var/gpdb/instance/datamaster/gpseg-1/gpperfmon/logs/gpmon.*.log gpperfmon logs
/var/gpdb/instance/datamirror/gpsegn/pg_log/*.csv mirror segment logs
/var/gpdb/instance/dataprimary/gpsegn/pg_log/*.csv primary segment logs
/var/log/messages Global Linux system messages

Use gplogfilter -t (--trouble) first to search the master log for messages beginning with ERROR:, FATAL:, or PANIC:. Messages beginning with WARNING may also provide useful information.

To search log files on the segment hosts, use the Greenplum gplogfilter utility with gpssh to connect to segment hosts from the master host. You can identify corresponding log entries in segment logs by the statement_id.

The log_rotation_age configuration parameter specifies when a new log file is automatically created while a database instance is running. By default, a new log file is created every day.