MAT headless mode

ParseHeapDump.sh

#!/bin/sh
#
# This script parses a heap dump.
#
# Usage: ParseHeapDump.sh <path/to/dump.hprof> [report]*
#
# The leak report has the id org.eclipse.mat.api:suspects
# The top component report has the id org.eclipse.mat.api:top_components
#

./MemoryAnalyzer -consolelog -application org.eclipse.mat.api.parse "$@"

suspect_report.sh

#!/bin/bash
#
# run with './suspect_report.sh path_to_heapdump'
#

nohup ./ParseHeapDump.sh $1 org.eclipse.mat.api:suspects org.eclipse.mat.api:overview org.eclipse.mat.api:top_components -vmargs -Xmx96g -XX:-UseGCOverheadLimit > mat_report.out &

Prerequisites

• MAT distribution
• Shell access
• At least the heap dump size RAM available

Note: The ParseHeapDump.sh script should be part of your MAT distribution. So to run a suspect report the second script and the MAT distribution is needed.

Run

First see if the heap size defined in the suspect_report.sh is right for your case. If so execute ./suspect_report.sh heap_dump

As a result an archive with the report in form of HTML files will be created. (Resulting files are usually located in the MAT directory if you have executed the application from that location).

Note: This will only create the suspects, overview and top components report. If you want other stuff or only one report (e.g. suspect) replace the report id from the suspect_report.sh script with only the desired report and execute directly from command line.