OpenNMS Update v1.38Dec 13, 2000, 20:03 (0 Talkback[s])
(Other stories by Shane O'Donnell)
Date: Wed, 13 Dec 2000 12:24:50 -0600 (CST)
Vol 1., Issue 38
In this week's installment... * Project Status + Release Notes + Interesting Web Resources + Coding Projects Underway * Bugs in Documentation * The Wish List * Afterthoughts...
So far, the feedback on the 0.4.0 "Testdrive" release has been good, but limited.
We've noted almost 1000 downloads to date (which we find encouraging) and numerous bug reports as well as emails back to me pointing out other "unanticipated behaviors".
If you have any feedback you can provide on the product, please drop me a line and let me know. Or fill out a bug report. Or email the "Discussion" list. Anywhere is fine with me, but at this stage, FEEDBACK IS CRITICAL!
On the list of more "major" bugs, we've already addressed the following oft reported bugs in CVS, and a 0.4.1 release should be en route by the end of the week (Watch this space!) The more significant ones include:
* No "calendar.xml" or "onetime.xml" files
FIX: Get these files from CVS or an earlier release.
* RPM Install throws all kinds of errors about not finding a JDK 1.3
FIX: If you are using the IBM JDK, our pattern match to verify the version was close, but not quite right. It's now fixed and new RPMs of version 0.4.1 will be available later this week.
* Confusing Install Process
FIX: We're working on better, cleaner documentation and the doc (including the Quick Start Guide) will be revved in 0.4.1 as well. Look for more info there, and they are available in CVS as well.
Interesting Web Resources:
As we get fired back up on building the distributed architecture, SOAP once again rears its head.
One of our main problems with SOAP (so far) is a distinct vacuum occuring everywhere you would expect there to be documentation. While there doesn't appear to be much underway to address this problem, Doug passed along an interesting link to a new SOAP book hitting the shelves. I haven't read it yet (but it's on my list of things to do--maybe), but if any of you have and can provide some feedback or a review, it would be appreciated.
The details can be found at:
To date, my favorite is still Brett McLaughlin's Java & XML published by O'Reilly (gods of all books technical)
For those of you interested in the source of this info, check
Coding Projects Underway:
Still wrapping our brains around the distributed stuff again. Efforts are underway, specifically around the servlet-based receivers for some of this stuff, but I haven't made it to the office yet this week to give you any more details than that (I've been on the road, not just lazy)...
* Maji Prelim Work -- Rick is a ctive on the "events" mailing list.
* MIB Compiler -- Things are looking good with the MICA stuff. Interesting question has been posed about JavaCC and what exactly is entailed with Metamata's intermediate "license".
* Distributed Architecture -- Weave is back amongst those of us with a door in that big hole in the back of our homes, and the contractors have left. Detailed design work is well underway.
Bugs in the Documentation
We are aware of a few points in which our documentation is either wrong, misleading, confusing, or just weak. We've addressed most of these, per your input. But I just wanted to call out a few points for clarification.
Bugzilla is being used to track bugs and enhancement requests as they are found and registered. We consider typos and documentation problems to be "bugs" as well, but we've set aside one specific bug-- Bug #100 -- to track those issues specifically. If you find one register it there. It helps to keep down the noise-level and administration overhead for typographical errors, and it keeps our bug counts manageable when our worst enemy is our own fat fingers. Thanks for your help in this pursuit.
Next, OpenNMS is building a big product. And a product that is capable of managing large IT infrastructures is going to be a fairly complex tool. Thus its configuration can get complex as well. We've done what we can to simplify the process (and my kudos to Steve for much of the high-level design that has gone into that effort). But we DO have a need to address the complexity, and we've attempted to address it with a barrage of documentation, not the least of which is the Quick Start guide.
I won't deny anybody the right to bitch if they want, but please don't complain about the complexity of the product if you obviously haven't read at least the Quick Start guide. This weapon in our documentation arsenal goes a long way toward helping you not only sort out what you need to do, but should provide some insights as to how the product is actually going to work. It's worth a read.
Remember, there is very little difference between someone who won't read and someone who can't.
The Wish List
Testing continues. Keep up the good work!
We've still got some other stuff that we'd like you to consider as places you might be able to help us along:
* Custom pollers -- You can now see how they are supposed to work (and do). And you've got some templates for how you can create your own. Modifying out TCP pollers to check new services is not brain-surgery, and should be an easy way for you to contribute, as well as solving needs for your own environment. Some things to consider might be: ssh, LDAP, Application-specific hooks.
* Migration Tools -- Have access to one of the network management frameworks proffered by the big boys? You can see their DB and configs and ours too. How about a series of Perl scripts to convert things over?
* Alternative User Interfaces -- Want to build a map? How about a Web interface? VT100? GTK? What's your bag?
* Wildcard -- What don't we have that you want/need?
First off - trip to Pittsburgh is off. So no Penguins jersey.
Many thanks for the answers to my "What's a Bruin?" question. Apparently, it's a male bear, a brown bear, and most importantly, the Brown University Interactive Language, a language for IBM 360 environment with a PL/1 syntax (I've actually done PL/1, for those of you that thought this was funny and obscure...it's really tragically sad).
The responses on maps were good and reasonably consistent. Most of you said "Maps suck", but there were some good arguments on behalf of maps, most of which argued that with maps you can at least manage multiple screens, which is impossible trying to manage from an event browser. Of course, there were the "they make management happy" reasons too, but for the most part, no real compelling arguments on the need for a topologically-organized map vs. another graphical representation. Many thanks to those of you that responded.
We've been getting some good press lately, and Network World has featured us a couple of times (once in an open source article, once in a Linux editorial), not to mention the rest of the press coverage around the Testdrive release. Consider all of this ammunition as you begin to lay the groundwork in your organizations for deploying OpenNMS to manage your infrastructure. Managers like to think that they can justify decisions, and articles in trade rags certainly help them along that path.