Overview:
This document outlines the process in which to enable rolling log rotation to manage memory allocation against growing log files.
Action:
Note: The log configuration file should be modified on the HEAD NODE ONLY:
- Open the gpudb_logger.conf file in a text-editor :
$vi /opt/gpudb/core/etc/gpudb_logger.conf
- Edit line #54 from:
54: log4cplus.rootLogger=INFO, ToConsole
to54: log4cplus.rootLogger=INFO, ToRollingFile
- Next, uncomment lines #127 to #135:
127: log4cplus.appender.ToRollingFile=log4cplus::RollingFileAppender
128: log4cplus.appender.ToRollingFile.File=../logs/gpudb-rolling.log
129: log4cplus.appender.ToRollingFile.Threshold=TRACE
130: log4cplus.appender.ToRollingFile.MaxFileSize=200 MB
131: log4cplus.appender.ToRollingFile.MaxBackupIndex=4
132: log4cplus.appender.ToRollingFile.Append=true
133: log4cplus.appender.ToRollingFile.ImmediateFlush=true
134: log4cplus.appender.ToRollingFile.layout=log4cplus::PatternLayout
135: log4cplus.appender.ToRollingFile.layout.ConversionPattern=%D{%Y-%m-%d %H:%M:%S.%q} %-5p (%i,%T,%-18t) %h %F:%L - %m%n - Save the file.
- Next, Kinetica services will need to be restarted on ALL nodes. First, stop all services on the Head node:
$/opt/gpudb/core/bin/gpudb stop all
Important: Wait for ALL Head Node services to be fully stopped before continuing - Stop services on all worker nodes:
$/opt/gpudb/core/bin/gpudb stop all
- Start the Host Manager service on all worker nodes:
$/opt/gpudb/core/bin/gpudb start-host-manager
- Start all services on the head node:
$/opt/gpudb/core/bin/gpudb start all
Should you have any questions or concerns, please visit our Official Documentation and Support Page.
1 Comments