nmon Analyser
 |
This is an Excel spreadsheet that takes an output file from Nigel's Monitor (nmon)and produces some nice graphs to aid in analysis and report writing. It also performs some additional analyses for ESS, EMC and FAStT subsystems. It requires Excel 2002 or later. |
Please note:
- The Analyser is not officially supported by IBM and no warrantee is given or implied. Use it at your own risk.
- No support is offered by IBM. Please post any questions you have to the forum (see link below)
- nmon analyser is the personal project of the author - Stephen Atkins.
New features for Version 3.3
- Support for new features in NMON12 for POWER6 and AIX V6.1
- Improved support for topasout and Linux
- Dynamic invokation of Windows SORT command for large files
- Automatic support for locales that use comma as a decimal separator
- New facility to allow merging of NMON files for trend analysis
- Automated creation of user-defined PivotChart
- Removal of SVCTimes option (replaced by new sheet from NMON12)
- Removal of DISKSORT option - to see if anyone complains and in preparation for Analyser V4!
- Changed REPROC option to work in all multi-file situations
- New summary workbook for multi-file mode
NOTE: V3.3 is designed for use with NMON12 but will work with most older versions. In case of problems analysing older files, try using V3.2.7 (link below)
V3.3.4 addresses numerous issues:
- Files with timestamps > 5 characters in length can now be merged successfully. FIRST/LAST processing will still not work correctly. Note the limit of 65300 lines per section has not changed.
- Fixed a bug which occured when the TOP section contained between 65300 and 65535 lines.
- Fixed the formatting for the BBBF sheet.
- Fixed a bug on the MEMUSE sheet (data being overwritten)
- Changed the order of the series on the MEMNEW chart so that system% now appears at the bottom.
- Sheets with no charts are no longer positioned to the end of the data set
- Improved handling of MEMPAGES sheets collection
- Removed DISKRESP from documentation (moved to V3.4)
Instructions for use
- Make sure you have the latest version of NMON
- Use NMON to capture data to a file (using one of the -f, -x, or -X options). This will create a file in the current directory with a name made up from
the hostname, the date, the time and with a filetype of nmon. This is a comma-delimited text file with the data in snapshot order.
I recommend that you keep the number of snapshots to around 300 in order to avoid problems with over-large files. The graphs will also look better.
- If the file contains more than 65K lines you will need to sort it using the "sort -A" command:
e.g. sort -A filename.nmon > filename.csv
- FTP the file to your PC (using the ASCI or TEXT options)
- Now open the nmon_analyser spreadsheet, click on the "Analyse nmon data" button and select the .nmon file
Known problems
- Please make sure that you send the nmon files to your PC as TEXT. This will generate CRLF line endings (instead of just CR). Failure to do this will cause processing problems if the nmon file has >65K lines.
- The new merge feature will fail if the files have been sorted.
- If you use the NOLIST=DELETE option you must have LIST=SYS_SUMM,DISK_SUMM,CPU_SUMM as a minimum.
- The WINDOWS sort command will fail if the filename or the directory name contains blanks or special characters.
- FIRST/LAST interval processing does not work correctly for files that do not have T0001 as the first interval or which have timestamps > 5 characters in length
Next release
The next release will be V3.4. Currently planned:
- Fix the problem with FIRST/LAST interval processing for files that do not start with T0001
- Allow FIRST/LAST to be specified as time values rather than interval numbers
- Produce disk response times analyses
- Reduce the memory footprint to allow processing of larger files
- Remove more section dependencies
Links
