freebsd/contrib/ntp/scripts/stats

Statistics processing scripts (README)

This directory contains a number of scripts for use with the filegen
facility. Those files ending in .awk are for the Unix awk utility, while
those ending in .sh are for the csh utility. Normally, the summary.sh
script is called from a cron job once per day. This script processes the
daily loopstats, peerstats and clockstats files produced by the daemon,
updates the loop_summary, peer_summary and clock_summary archive files,
and deletes the daily files.

In the case of the Austron 2201A GPS receiver, the clockstats file
contains a wealth of additional monitoring data. These data are summarized
and writted to the clock_summary file, then a series of special files are
constructed for later processing by the S utility.

The summary.sh script invokes a number of awk scripts to actually produce
the data. This may result in multiple scans of the same input file.
The input file is deleted after processing. In fact, the shell scripts will
process all input files found of the correct type in chronological order,
deleting each one as it is scanned, except the current day file.

The summary.sh script can produce input files for the S utility, if it
is found on the search path. This utility makes PostScript graphs of the
loopstats data for each day, as well as various statistics produced by
the Austorn 220aA GPS receiver. The S utility is automatically run
as a background job. Its control files have the .S extension.

The psummary.awk script can be used to scan the peer_summary file and
construct an historical reprise of the daily summaries. 

The file formats are documented in the README.stats file and in the
scripts themselves. Further detail on the radio clock ASCII timecode
formats and related data are in the README.timecode file.

David L. Mills
University of Delaware
mills@udel.edu
1 November 1993
Revised 12 April 1994