/*
** SellaNMS
** $Id: TODO,v 2.66 2006/01/21 23:35:03 sella Exp $
** Copyright (c) 2001-2005 Digital Genesis Software. All Rights Reserved.
** Released under the GPL Version 2 License.
** http://www.digitalgenesis.com
*/

ToDo List
---------

To see updated information visit http://bugzilla.digitalgenesis.com.

[Show Stoppers]

Fixup the INSTALL document. Its heavily wrong.

test dsc.snmpd with bad and multiple communities.

Finish Administrator's Manual.

[Nice to Have]

Move query in map statements into then clause as "query database".

Write a developers manual that covers writting new modules and the use of the primary libraries.

Get the application to compile with concept checks enabled.

Document somewhere "export MALLOC_CHECK_=2" to get more info from malloc() backtraces.

Document somewhere "export LD_ASSUME_KERNEL=2.4.1" to disable NPTL threads.

Document somewhere "export GLIBCPP_FORCE_NEW=1" and "GLIBCXX_FORCE_NEW=1" to disable arena memory.

Verify Arbor trap:  Nov 7 00:01 12.101.21.12 Report @.1.3.6.1.4.1.9694.1.4.1.16@ (id: @.1.3.6.1.4.1.9694.1.4.1.17@) completed.

Verify Arbor trap:  Nov 7 11:25 12.101.21.12 BGP session with CHCG-RTRX-02 has bounced @.1.3.6.1.4.1.9694.1.4.1.7@ times.

Verify Arbor trap:  BGP Trap @.1.3.6.1.4.1.9694.1.4.1.8@ for event @.1.3.6.1.4.1.9694.1.4.1.9@ on prefix: @.1.3.6.1.4.1.9694.1.4.1.10@

Verify Arbor trap:  Autoclassification started on collector @.1.3.6.1.4.1.9694.1.4.1.1@ by @.1.3.6.1.4.1.9694.1.4.1.22@

Auto-discovery via incoming SNMPTrap, Syslog, etc. If a host sends a trap with the correct incoming community, add that IP as a location to poll during the next discovery. This would work out very well for companies that do managed routers.

Verify that snmptrap proxy works (specifically v2). The source address becomes the proxy (RFC 2576 Sec 3.1 talks about how to handle v1->v2 trap proxy, but I can't find a source covering straight v2 proxy)

Add SMTP auth support to email module.

Add functionality to state daemon to alarms on bouncing interfaces.

Rewrite history duplicate suppression to be an in memory algo instead of polling the database for each incoming alarm. Its way too slow currently when the database grows.


[Ideas]

Use input policies up front as information is recieved from the network (in monitoring modules) and output policies for only processing output. This will decrease the amount of data that needs passed though queues in the module. As long as input policies are simple and don't use maps it shouldn't bottleneck.

If guid gets a duplicate event (matching state), then it could increment a duplicate counter, rather than display it twice or remove the previous alarm. This could be implemented pretty easily as an action (like clear).

Example GUI Reports

Add backup SQL server support to PHP GUI. (ie: First is not avail, fail over to a backup)

Look into a knob to enable auto-generating CPE devices which connect to edge interfaces. Assign the other end of a /30 to the CPE interface allowing the NMS to monitor the CPE device.

SNMP module capable of doing on-demand polls of BGP neighbors, interface up to clear traps.

SNMP module capable of generating alarms based on interface drops or over X% util on a link.

SNMP module that examines interface drops and generates a trap. Could be correlated to packet loss events.

Give users ability to view previous discoveries. Would require changes to the db structure.

Allow trace-options to use a policy to filter/limit output. Its origin would be "trace-options".


[Policy Additions]

Syslog policies for the following:

M160:
Add errors message about SFM failures.
