OpenNMS Update v1.35
Nov 22, 2000, 06:12 (0 Talkback[s])
(Other stories by Shane O'Donnell)
Re-Imagining Linux Platforms to Meet the Needs of Cloud Service Providers
Date: Tue, 21 Nov 2000 18:24:48 -0600 (CST)
Subject: [opennms-announce] OpenNMS Update v1.35
Vol 1., Issue 35
In this week's installment...
* Project Status
+ Brrr...Code Freeze
+ More Stress Test Results
+ Documentation Available
+ Coding Projects Underway
* OpenNMS & Giving Thanks
* The Whitepaper: Thanks for your Feedback
* The Wish List
We're right on the brink of reaching one of our major milestones. As
of tomorrow, we are putting a temporary hold on new development to get
our current code base cleaned up, prettied up, and releasable.
This milestone is in support of our stated goal to provide a release
by the end of November that will provide significant functionality,
albeit limited by a lack of distributed architecture. Internally, we
are referring to this as our "Testdrive" release, in that you'll be
able to use most of the major features of the product including
discovery, capabilities checking, adding nodes to the database,
service polling, event handling, automated actions, integration with
trouble-ticketing/notification systems, and graphical administrator
and operator interfaces.
In reality, the "Testdrive" release will be all some smaller networks
will ever need, and price aside, the current feature set already
compares favorably with some of the commercial "closed source"
We are currently at stable release OpenNMS-0.3.0 and development
release OpenNMS-0.3.1 (available via CVS). Once we've frozen the code
for new development and gone through an iteration of testing and bug
fixes, we'll release (as stable) OpenNMS-0.4.0. When this release is
available, it will also be available from our "Downloads" page as
And for those of you grabbing code and doing your own homegrown
testing, don't forget to register your bugs with us at
More Stress Test Results:
Last week I reported some early results in some very preliminary
stress testing. Let's just suffice it to say that I hit the "Send"
button about 45 minutes too early...
Not too long after I sent the Update last week, Mike came sauntering
into my office with what he considers somewhat more "real" numbers.
Before I report the more recent findings, keep in mind that our "test"
included receiving SNMP traps on the same interface as we were
updating a remote database (which maintained our table of events).
Every trap that we received was converted to an internal event, mapped
with configured fields, passed through the system, handed to actiond
(to run the automated actions), and written to the database.
With all this going on, we were able to handle close to 300
traps/second without losing data (and when data loss started, it was
likely due to interface congestion).
That's right. 300. 3-0-0. Three Hundred. 100x3. A lot.
A very early draft of an Administrator's Manual has been built and is
available electronically via CVS.
Big thanks go to Steve for knocking this one out. Documentation is a
If you want a copy of the doc, the best thing to do is to do a
complete CVS checkout (e.g., cvs co bluebird-all) and then build
the full project (e.g., build.sh all). This way, you'll get the latest
version of the software as well as the documentation built.
As a word of warning, this may take a while. On a P3-600 with 160MB of
RAM, the project itself will build in less than thirty seconds. If you
build the documentation too, it's over 10 minutes, so plan your coffee
Coding Projects Underway:
* create.sql -- No changes this week.
* Events -- More stress-testing results above.
* SCM -- Everything is at least partially integrated with SCM
* OutageManager -- Reviewed and committed to CVS for initial
* RTC -- Still working on this one. The Real-Time Console works, but
now we're trying to iron out some "Swingy" things.
* Filters -- Solid. Now we're looking at how the interfaces build
* Maji Prelim Work -- Rick is active on the "events" mailing list.
* joeSNMP -- Kicks ass, takes names.
* Operator Panels -- Much tweaking and bug fixing here, not too
mention all of Jacinta's RTC work.
* Admin Panels -- More bug fixes. Looking good.
* Documentation -- Administrator Manual available via CVS.
* Installation -- Working on building RPMs for Linux platforms.
Currently have working spec files, so when the "Testdrive" release
is done, look for easy-to-install Linux options.
OpenNMS & Giving Thanks
Now is that time of the year when we pause for a moment, look around
us, and note how lucky we are.
From a project perspective, we've got a lot to be thankful for:
* A Great (and Growing!) Community of brilliant minds helping our
* The Atipa Corporation, for smokin' Linux boxes and paychecks
* Sun & IBM, both making Java viable on many platforms
* Just about the whole damned Apache project. You guys rock!
* The U.S. Constitution, for all this entertainment
* Linus Torvalds and Alan Cox. 'nuff said.
* The Creator of the Universe, for caffeine in all naturally
* Science, for caffeine in all non-naturally occuring forms.
* Nipsy Russell
* Tobias Oetiker, Dave Rand, Jim Trocki, Mark Spencer, Jeff Allen,
Juergen Schoenwalder, Jorgen Frederich, and all of the other folks
making strides building open source network management tools.
* Weave, Sowmya, Steve, Mike, Jason, Jacinta, Ben, and Vishwa for
making this a cool place to hang out for 12 hours a day
* And you, for not only tolerating my weekly rants, but for reading
this far in this one.
The Whitepaper: Thanks for your Feedback
Just a note of thanks to everyone who downloaded and read the
whitepaper I offered a couple weeks ago. Much to my surprise, there
was a whole lot of interest in this one, and people are evidently
reading it, since a lot of the feedback I've been getting is on typos
and grammatical problems.
For those of you that didn't get it, it's still available for download
from the website at:
Thanks again for your interest and feedback!
The Wish List
As we close in on some critical dates, we need your help more than
ever on some of these key project components. TIA!
* Testing of the entire product, with bugs reported to Bugzilla
* Add SNMPv3 functionality to jSNMP
* Implement the event correlation spec.
* Consider the integration points for CIM/WBEM support
* Documentation (both lightweight and detailed)
* Send a note to IBM's alphaWorks site and ask them to open source
the ASN.1-to-XML converter from the XML Security Suite!
Just as a "heads up" to those of you paying attention, starting
immediately after Thanksgiving, we're going to dedicate a small subset
of the core team to architecting and building out a client-server
poller plug-in that will essentially set us up to do agent-based
management where appropriate.
We've already gotten a lot of good input on this, from sources you
wouldn't necessarily expect: individuals wanting to do storage
management, commercial organizations wanting an open source solution
to manage their systems (it's cheaper than the existing systems
management solutions!), and other open source efforts that have built
Our approach is pretty simple. Write a simple Java-based plugin that
fits into the poller architecture and provide some well-defined
messages that the server can send and receive. The rest of the
back-end services we've already got (for the most part) and the rest
of the client services are likely platform/application/system-dependent.
Any thoughts on this one? Drop me a line. And thanks!
My apologies to Soupy Sales for leaving him off the list,