Introduction
PIKT is a cross-categorical, multi-purpose toolkit with uses limited only by your imagination.
Here are several case studies, with examples and commentary, introducing some of PIKT's varied uses, and many of its special features:
-
Reporting a Problem (high system load average)
-
Fixing a Problem (kill idle user sessions; monitoring user activity)
-
Reporting and Fixing a Problem (delete junk files; disk space management)
-
Reporting Multiple Problems (service failures; system downages)
-
Scanning a Log File (dmesg scan; log file analysis)
-
Scanning Multiple Log Files (syslog scans; PIKT log scans)
-
Site-Wide System Scanning (interactive system monitoring)
-
Enhancing the Command Line (remote command execution; command line macros)
-
Monitoring System Security (checksum differences; change auditing)
-
Job Scheduling (centrally directed scheduling daemon; cron alternative)
-
Monitoring SNMP (air conditioning & power failures; monitoring SNMP-enabled devices)
-
Configuring a System (nsswitch.conf, motd; system and application configuration)
-
Configuring a Network (resolv.conf, named configuration files; DNS, BIND, and network configuration)
-
Reporting Policies (alert e-mail routing and scheduling)
-
Script Development and Testing (debugging, testing; private vs. public reporting)
Following is a paper presented at the LISA 2000 Conference December 2000 in New Orleans. It is a good, concise, and still accurate if increasingly outdated introduction to PIKT.
- PIKT: Problem Informant/Killer Tool, Postscript version (gzip'ed).