Skip to content | Change text size

ITS home

 

About MRTG and NetWare

These pages are created using the MRTG graphing package and a helper application called MRTGEXT that provides NetWare specific information.

See Also

Overview Page

The server overview and volume statistics pages are generated by a modified version of the scripts provided by James Drews, which is distributed with MRTGEXT.

The role account cruxnds on the server webedit is used to house and run these scripts.

The package is installed in ~cruxnds/pkg/nw-mrtg and is run every half hour by a cron script on webedit.monash.edu (ref. man crontab(1)).

At 18 and 48 minutes past the hour, the server status script ~cruxnds/pkg/nw-mrtg/servstat/servstat.pl is run. The list of servers to check, in fully-qualified DNS form, is contained in the configuration file servstat.cfg.

At 3 and 33 minutes past the hour, the volume status script ~cruxnds/pkg/nw-mrtg/volstat/volstat.pl is run. The list of servers to check, in fully-qualified DNS form, is contained in the configuration file volstat.cfg.

Both these scripts have been modified to write only HTML fragments, these are included by server side includes into the ITS webpages.

Detail Pages

This is a mixture of dynamic pages, static pages and dynamic graphs. The dynamic pages are used for server status summaries. These take query strings or variables passed from a form to display specific server information.

All server utilisation pages are called <server-name>.1.html, all server connection count pages are called <server-name>.2.html.

MRTG is installed in ~cruxnds/bin and invoked every five minutes by a cron script on webedit.monash.edu. The MRTG configuration file (mrtg-webedit.cfg) and data files are in the directory ~cruxnds/lib/mrtg2/monash, while output files are all kept in /web/its/novell/monitor.

The file mrtg-webedit.cfg specifies all the server attributes that will be logged, it is laid out in sections so that defaults for a particular attribute are specified (such as headers, maximum values), then each server for which that attribute will be monitored is listed, then the defaults are cleared. Please read this file in conjunction with the MRTG manuals. A number of headers and proformas are used so that the generated pages match the rest of the ITS website in look and feel.

TODO

  • Replace the static index pages with a dynamically generated index, so that additional servers or attributes can be easily added.

BUGS

  • Due to the large number of servers being monitored, there can be considerable delay in receving responses to the polls. When several servers are uncontactable for an extended period, the processes will stall and the number of processes owned by the account cruxnds rapidly approaches the maximum allowed by the system. When this maximum is reached, the entire data-gathering system stops and the only solution is for an administrative user on the webedit system to kill of all processes owned by cruxnds.