|
|||||
|
I’ve gotten all of the dependencies of OpenNMS packaged up on Fedora Core 8 now, and have our yum repository up-to-date. If you’re looking to install on FC8, you should be able to follow the usual instructions (substituting “fc8” for the distro) and it should work fine. Share on FacebookFigured it’d been a while since I did a general “status” post about what’s been going on. First and foremost, as I mentioned before, much of my spare time has gone into fixing up the Fink package database. It’s now much easier on our web server, and uses a combination of PHP and a really spiffy Lucene-based full-text search engine called “Solr.” Also, I did some work on making Fink play nicer with the new Xquartz releases. It’s still in testing, but in the meantime, their 2.1.1 release provides a workaround to allow Fink users on Leopard to run without issue. I’ve also been wearing my OpenNMS.com hat recently, and am working on some spiffy customer-management tools for keeping track of our support work better. I’m a big fan of Ruby on Rails, and am mocking it up in that — on top of JRuby, of course. In the process, of getting my development system set up to do rails development, I updated all of the Fink packages for rails 2.0 and related stuff, as well as taking over rubygems from . . . → Read More: It’s Been a Crazy Couple of Weeks Share on FacebookOpenNMS 1.3.9 has been released. It’s mostly a few small bug fixes and features, but it does fix a rather important deadlock in the database code that could cause notifications to not go out correctly (among other things). Other than that, we’ve been pow-wowing on what to do for the next release — there’s a bunch of niggly little bugs that need to be cleaned up, but overall, the trunk codebase is still looking solid. Share on FacebookWe’ve finally gotten the OpenNMS image for VMware put together and uploaded to SourceForge, with OpenNMS 1.3.8 on it. It hasn’t been updated since the 1.3.3 VMware image was put together. Since I’d been having such good luck with Mandriva, and as Tarus mentioned, they’ve been very supportive of us in general, I went ahead and made the image using Mandriva 2008.0. It has the advantage of being quite a bit smaller than our previous CentOS-based image (about 400MB compressed) and is a very minimal install with just the minimum needed to get OpenNMS up and running. So if you were wanting to try OpenNMS out but didn’t really have anywhere to run it, why not download the image from SourceForge and give it a shot? All you need is the free VMware player; just follow the README file inside the package and you should be off and running. Share on FacebookI’ve finished getting everything set up so that you can run OpenNMS on Mandriva Linux. It is now possible to install using URPMI with a minimum of fuss. I’ve gotta say, it’s been a number of years since I’ve tried out Mandrake Mandriva, and it’s definitely grown from just a KDE-themed RedHat into a pretty sweet and polished Linux distribution. Share on FacebookOpenNMS 1.3.8 is out. The biggest change is that it runs on Windows, with a spiffy GUI installer, even. Well, technically the installer should work for any platform, but it was written to make installation easier on Windows, where there is no real package management. 😉 RPM and Debian packages are prepared as well, so everything should be good to go. There’s plenty of other nice small changes in this release across the board; it’s definitely clear that we’re approaching ready to put out a stable release. Other than a few small features that we still want to implement, things are definitely looking really solid from a bug standpoint. (A sizeable percentage of our support customers are already running on the 1.3.x series, it’s so much better than 1.2.) So if you haven’t tried OpenNMS out yet, check it out, it’s the best release yet! And feel free to drop by the discussion lists or irc if you have any issues. Share on FacebookSo I’ve spent the last week or so working on a Windows installer for OpenNMS, using IzPack, an awesome Java-based installer. After some trial-and-error getting it to handle paths nicely (some of our code is not spaces-in-paths clean, so I had to hack something up to get the DOS 8.3 filename), it seems to be working! We’re going to spend some time testing it and making sure everything works well enough to be considered an “alpha”, but this certainly appears to put us on track to have a nice installer working shortly after 1.3.8 is out — which should be any day, there are only a couple of bugs left on the blocker list. …and since IzPack is a Java-based installer, it actually works on Mac OS X and Linux as well, and in theory, anywhere else that supports Java 1.5… (Although it is still recommended you use native package management, since you’ll get config-file management and other nice stuff…) Share on FacebookLet me start with a story. It’s a story of answering a simple question in the #fink irc channel. What was my answer? It doesn’t really matter. What matters is that I used “..\lib” in my response. Why does it matter? OH GOD I JUST ACCIDENTALLY USED A BACKSLASH TO REPRESENT A PATH IN IRC! That’s right, this week I’ve been in the Land Of Evil, working on porting OpenNMS to Windows. I’ve been so heads-down into it, I actually started thinking in backslashes even in a Mac OS X channel. Oh, the shame. <grin> Anyways, it actually (surprisingly!) mostly works. The hardest part was porting jicmp, which required setting up a mingw environment and fixing our configure stuff in a lot of ways. And I’ve gotta say, libtool and I have had our differences in the past, but it performed beautifully at hiding the details of making a .dll file out of our code. There’s still plenty left to do. The docbook stuff doesn’t run right. Our GWT maven plugin inexplicably fails, even though the command-line it generates actually works . . . → Read More: Embracing and Extending OpenNMS Share on FacebookFirst of all, I want to point out that OpenNMS 1.3.7 is out, and it’s full of awesomeness. Big speed improvements, my SmokePing clone, and lots of other cool stuff. I finally finished up the last of the packaging updates today; 1.3.7 is now in Fink, has Debian packages, and RPMs. I also spent a lot of time updating the installation instructions (Yum, Debian) so please, try it out, and if you run into any issues, let me know, and I’ll make sure the docs get fixed. In other news, I’ve actually started spending some time getting Fink stuff up-to-date again. PostgreSQL has bugfix releases coming up for all supported releases, and I have some KDE updates coming as well. In addition, I need to catch up on the KDE/Mac stuff, I’m going to start working on a new build this week. Share on FacebookSo I know I’ve been quiet, but I’ve been busy as a bee, mostly with OpenNMS. First of all, I’ve been spending time prepping for the 1.3.7 release, closing bugs, documenting, and writing code. 1.3.7 is ramping up to being an awesome release. First of all, brozow found a pretty major performance issue in the way we’re scheduling and writing RRD data. It is now an order of magnitude faster to collect the data and get it queued for writing, so once again we’re I/O bound. 😉 Second, we spent some time completely reworking the ICMP (ping) code, it is crazy efficient now, and the code is considerably easier to understand. No more 2-pages-of-code-to-do-a-simple-ping. Discovery is way faster, and overall ICMP handling is much improved. That paved way to the new MultiICMP monitor, a clone of Tobi Oetiker’s wonderful SmokePing tool. I wrapped that up this week, and it’s happily chugging away in the new codebase. Also, we’ve embedded Jetty into OpenNMS, so we no longer require Tomcat for our Web UI (although we still offer the option if . . . → Read More: OpenNMS Rockin’ Share on Facebook |
|||||
|
Copyright © 2024 Tales of the Raccoon Fink - All Rights Reserved |
|||||
Comments are closed 
