NodeWatch


NodeWatch Documentation

How NodeWatch Works

NodeWatch Status

NodeWatch.Nodes

NodeWatch.Actions

Qpage.CF

NodeWatch.Periods

NodeWatch.Options

NodeWatch in Syslog


NodeWatch Documentation

NodeWatch is a public domain PERL package designed and written by the Administrative Division's Information Technologies Department at the Fred Hutchinson Cancer Research Center to monitor InfoTech's portion of the Hutch data network. This page describes how NodeWatch behaves and how InfoTech has configured it to support its efforts at monitoring aspects of the Hutch data network environment. See How NodeWatch Works for a discussion of the NodeWatch package and how it functions.

NodeWatch monitors a list of IP-accessible nodes via the ICMP echo protocol, flagging them as either up, down, or unknown. "Up" nodes are nodes which have been responding to pings for some amount of time. "Down" nodes are nodes which have been missing pings for some amount of time. And "unknown" nodes are nodes which NodeWatch either cannot resolve (cannot convert the host name into an IP address) or nodes which NodeWatch hasn't been pinging for long enough to make a determination as to their status. See NodeWatch Status for a screen dump illustrating the Web interface to NodeWatch.

To analyze how NodeWatch is currently configured, have a look at its configuration files. Collectively, these files describe how NodeWatch behaves in the InfoTech environment. These files are copied to this Web page every night; therefore, the files you see here are typically hours stale from the actual files supporting InfoTech's live NodeWatch installation.

  • NodeWatch.Nodes: To see the list of nodes which NodeWatch monitors, see the NodeWatch.Nodes file.

  • NodeWatch.Actions: To see the list of actions which NodeWatch can perform, see the NodeWatch.Actions file.

  • Qpage.CF: NodeWatch's actions are simply command-line binaries with the appropriate node transition information handed to them. The examples posted here employ Qpage, a public domain SNPP (Simple Network Paging Protocol) utility. To see the groups defined in Qpage, as they relate to nodewatch.actions, see the Qpage.CF file.

  • NodeWatch.Periods: To see the list of time periods which modify whether or not NodeWatch performs an action (sends a page), see the NodeWatch.Periods file.

  • NodeWatch.Options: To see the global potions which control NodeWatch's overall behavior, see the NodeWatch.Options file.

  • NodeWatch in Syslog: This page illustrates a typical day of NodeWatch activity, as seen in syslog. These extracts identify the times when NodeWatch missed pings to monitored nodes.


Prepared by:
Stuart Kendrick

Last modified: 10-Oct-2004