Network Mgmt |
Problem Mgmt |
I am particularly interested in network management, which I define as the set of practices which support our efforts at providing reliable, responsive packet transport. On these pages, I offer tools for device management (configuring the electronics which push packets from source to destination), fault management (identifying when one of those electronics is misbehaving, performed here by polling and trapping and thresholding), and trending (looking at the past in an attempt to predict the future, i.e. capacity planning).
When I am performing ad hoc queries against an SNMP agent, I rely on the Netops-MIBs collection of MIB files, along with the Net-SNMP Toolkit and this snmp.conf file, to translate between OIDs and Object Values. I've been accumulating these MIB files for years; if you have a favoriate MIB file which you suspect I'm missing, please send a copy to me. Periodically, I download the latest Cisco MIB collection, fix syntax errors, and integrate them into the Netops collection.
Generically, these scripts support options to parse a hosts table or a file, extracting targets of interest based on regular expressions, a "dry run" mode in which they gather information but don't actually make changes, and the ability to run either interactively or via cron.
If you have a favorite manufacturer/model which these scripts don't support and are able to give me SNMP access to a sample box, drop me a note -- schedule permitting, I'm willing to extend these scripts to make them more comprehensive.
We use Qpage to propagate notifications to pagers, cell phones, and other portable devices, consulting TAP Paging Terminal Phone Numbers to define a service provider and pointing techs at BrokeBit's description of this strategy for illumination. We maintain a telnet interface for techs to use when paging each other and employ Perl code to automate sending paging from inside applications.
Many of these scripts can automatically generate a list of relevant devices (switches, routers, UPSes, etc.) by parsing an /etc/hosts file ... yes, how quaint! ... we use Zone-to-Hosts to perform a zone transfer from a DNS server, parse the result, and produce an /etc/hosts file.
Console software delivers a view into the state of networked resources, in our case, bringing together the view of multiple polling and trapping applications into one screen.
Device management offers links to in-house code which we use to manage our flock of network-oriented devices, from switches and routers to UPSes and Environmental Monitoring Units.
Log Analysis offers links to in-house code which poke through syslog and produce reports.
Polling offers links to in-house code which we use to poll devices and servies in an effort to identify when they have quit functioning.
Trapping offers links to in-house code which we use to watch syslog for signs of device instability.
Trending offers links to in-house code which we use to trend resource consumption on network devices and end-stations.
Internet links to sites which tackle these or larger problems.
Last modified: 16-June-2012