Documentation/devtools/maemo5/sp-endurance-postproc

kaltsi (Talk | contribs)
(New page: == Description == ''sp-endurance-postproc'' provides post-processing tools for producing report(s) from the data collected by sp-endurance. ...)
Newer edit →

Revision as of 06:48, 1 October 2009

Contents

Description

sp-endurance-postproc provides post-processing tools for producing report(s) from the data collected by sp-endurance.

These include:

  • a script for producing an overview HTML report of the collected endurance data
  • a script for splitting the data into several reports (in case of reboots during automated testing)
  • a script for for parsing errors and other significant events from syslogs

Packages

source: sp-endurance

binary: sp-endurance-postproc

Usage Examples

Once you've transferred the endurance data saved by [sp-endurance] to a PC, you can post-process it with the following:

# apt-get install sp-endurance-postproc
# parse-endurance-measurements test-case-name/[0-9]*

This will produce you a nice (single file) HTML report which you can send in E-mail, attach to Bugzilla etc.

The report lists first the information about the system's intial state (first measurement data), then the system's memory usage change overview, more detailed memory usage of each process whose memory usage changed and last details about all resource usage changes and (recognized) syslogged errors between each of the measurements. Finally there's a summary of the differences between the second and the last measurement.

If you want to get a similar report for a subset of these measurements (e.g. because there was a reboot between them or to see the leaks better you want to skip some of them), the script offers several options for that too.

Other platforms

If you want to do the post-processing on platforms like Windows which may have Python, but not Bash, you can call this directly:

# endurance_report.py test-case-name/[0-9]* > report.html

Syslog parsing

As a part of endurance data processing the saved syslogs are parsed for errors and other significant events, but you can do that also directly with:

# syslog_parse.py syslog.gz

The syslog file can be compressed and you can get the lists either in ASCII (default) or in HTML.

Links

Man pages:

[endurance_report.py](/development/documentation/man_pages/endurance_report_py.html)

[parse-endurance-measurements](/development/documentation/man_pages/parse-endurance-measurements.html)

[split-endurance-measurements](/development/documentation/man_pages/split-endurance-measurements.html)

[syslog_parse.py](/development/documentation/man_pages/syslog_parse_py.html)

See Also

sp-endurance, syslog