Earthquake in Chile

One of the major things I love about OpenNMS is its worldwide scope. Last fall we had two students, Javier and Cristian, attend our training course in Pittsboro. They work for Telmex in Chile, and the office is just outside of Santiago.

When I heard about the magnitude 8.8 quake, my first thought was to hope that they were okay. I did get an e-mail from them and they are fine, although Cristian writes “I am hoping the earth stops moving … it is a worrying feeling all day long to feel the floor moving and not knowing when it will come.”

Our best thoughts go out to everyone affected by this earthquake, as well as our hopes for a speedy recovery.

Open Source and Building Networks

Geoff Davis, a friend of mine from high school who now works at Google, wrote a post about how open source software projects can be used to build up a social network.

I sometimes get jealous of college kids today. When I was in school, computer networking was pretty much limited to BBSs and 2400 baud modems (although the college did have Internet access). I wonder what would have happened to me if open source had be prevalent in those days. All the time I spent drinking beer could have been turned into something more productive. (grin)

As Geoff points out, not only does working on open source projects give you something to put on your resumé, it allows you to make the connections to get that job in the first place.

The Linux Link Tech Show

Okay, I’m scared. I was recently invited to participate on The Linux Link Tech Show (check it out on Wednesday, 3 March). I said, sure, but that was before I actually listened to an episode. It’s pretty wide open.

At least I should be able to deal with the format, since each show is over 2+ hours long. I once applied to participate in a 5 minute lightening talk and the guys at the office just laughed, saying that it takes me more than 5 minutes to say my name.

Also, there doesn’t seem to be many limitations on language. I got bleeped on FLOSS for a rather minor vulgarity, so it will be interesting to see if I can keep it clean.

Anyway, if you have a couple of hours to kill next Wednesday, check it out.

OpenNMS in Africa

For over three years now, OpenNMS has been heavily involved in bringing open source development techniques to the world’s largest telecommunications providers through the TeleManagement Forum (TMForum).

Dr. Craig Gallen (OGP) is the leader of the TMForum Interface Program (TIP) and he is making a number of presentations about our work with TIP. The first one is being held at the Management World Africa conference tomorrow (25 February).

For many years carriers have be asking for a greater number of open interfaces so that the various management products they need can more easily interact. Needless to say, getting proprietary software companies to share their work has been difficult, but the hope is that by using open source techniques along with permissive licensing we can both increase the number of open interfaces as well as speed their development.

Craig will also be at the Management World Middle East conference, and both of us will be at the main Management World conference in Nice, France in May. If you are going to be at any of these, please stop by and introduce yourself, and we can explain in more depth what we are trying to accomplish.

SCaLE 8x

I have safely made it back home from SCaLE 8x. Once again it was a great show. Unfortunately, I didn’t get to see many talks since I was manning the OpenNMS booth, but the feedback I hear was that most of them were excellent. I really hate the fact that I missed seeing Brian Aker, although he posted some interesting comments on his blog about the show.

My Sunday keynote seemed to be well received. I was a little nervous – not because I was speaking in front of a large crowd, but because I like the show so much I didn’t want to disappoint. If you were unable to see it and are curious about my thoughts on starting an open source business, it is available via streaming on-line.

I made some new friends and after the show managed to spend some time with Karen Sandler and Bradley Kuhn. Bradley and I were able to continue our debate on copyright assignment, with me firmly in the camp of dual copyright and him warning that the potential for evil is still too high. I don’t think we changed each others mind any, but like any rational argument I hope that we each can see better into the others reasoning.

I always love the fact that user group and volunteer driven conferences like SCaLE seem to be better than the “professional” conferences I attend. I am looking forward to SELF in June, where we are a diamond sponsor. Jeff went last year and said it was great, but this will be my first time attending.

More importantly, I am getting more and more excited every day about the OpenNMS users conference in May. I couldn’t be happier with the speakers. The first day, with the exception of my talks, is completely made up of users of OpenNMS who can tell their stories of how they use it, why they like it, and how it makes their jobs easier.

The second day features some of the main developers of the project giving in-depth workshops on the internals of OpenNMS. If you ever wanted to get started customizing the software, this is the place to be.

Remember, early-bird registration is available at a discount and we are required to limit the number of attendees due to space constraints, so if you are interested in coming be sure to register early and often.

Hope to see you there, at SCaLE or at SELF. Be sure to stop by and say “hi”.

NASA-SE Winter Meltdown at CMP

Jason has posted an update for the first race of the OpenNMS Spec E30 BMW.

Notable takeaways:

  • Personal best lap of 1:54.2, beating the previous time by 1.5 seconds
  • … glued to Brian’s bumper for the next two laps before executing a spectacular 360 degree spin in the kink … finished tenth but lowered personal best to 1:53.3
  • made one last pass on the final lap and held Steven off for eighth place

Unfortunately, his camera rig failed to work so there is no video of Sunday. Next week is Road Atlanta for a BMWCCA race.

Kevin Smith – Too Fat to Fly?

I saw a news article this weekend concerning the director/actor/writer Kevin Smith. It seems he was bumped from a Southwest Airlines flight from Oakland to Burbank because a Southwest employee determined he was too big to qualify to travel in just one seat.

He apparently used his powerful social networking presence to “scorch the earth” with his complaints.

The first thought that went my mind was, dude, I really like your movies and Chasing Amy is a classic, but as a frequent flyer, if you are too big to fit in a seat you should have to buy two tickets. Otherwise, you end up in part of my seat that I paid to use, small as it is.

But the second thought that went through my head, hot on the heels of the first, was, wait a second, Kevin is a big boy and I’ve never seen him in person, but he’s not that big. I’ve sat next to much larger people.

Perhaps the media is, err, making a bigger deal (forgive the pun) of this than is warranted?

I was catching up on my RSS feeds this morning and noticed that Kevin had posted on his blog (which is rare). In it he tells his side of the story, which is quite at odds with the published accounts, where it turns out he was flying standby and Southwest made the decision that they could not accommodate him. For some reason they used his size as the excuse to bump him from the flight.

What is really making him angry is that now he has to carry around this very public stigma of being “Too Fat to Fly”.

Look, as someone who has been overweight much of my life, I can sympathize. One of the reasons I work so hard is to overcome the stereotype that fat people are lazy. If you haven’t lived through it you can’t understand.

Now this doesn’t mean I’m forgiving of having an extra-large person sitting next to me on a flight. It’s not an issue of discrimination, it’s an issue of my rights. I paid for a certain amount of room and I should get to use all of it. If you need more room than the default, buy more.

However, Kevin met all the criteria for flying in one seat, and telling people otherwise is discrimination.

I’m a frequent flyer, and it is amazing how many frequent flyers have my approximate body shape. Stick three of us in the exit row of an MD-80 and, yes, we’re going to be a little cramped. We all fit into the seats but there isn’t always enough shoulder room. That’s just the way modern coach seats are laid out. Since it helps to have “elite” status from frequent flying to get in the exit row in the first place, you can imagine that this occurrence isn’t all that unusual.

But I think Kevin’s treatment is a symptom of a much greater problem in air travel. Somewhere in the last 20 years we’ve gone from honored guests to criminals.

First, there is the demoralizing security process. You might claim that it is for our own safety, but the fact of the matter is that it is security theatre. The process didn’t stop a man in the UK from trying to bring a liquid bomb on board, but it now requires every one of us to have our shampoo bottles scrutinized. A man flying into Detroit tried to set his undies on fire, and now every one of us has the potential to be photographed basically naked, even though it has been demonstrated that it is still possible for a determined person to get a weapon on a plane even with full body scans.

One of the most sane articles on the subject comes from the Salon “Ask the Pilot” column which states in part:

… over the five-year span between 1985 and 1989 we can count at least six high-profile terrorist attacks against commercial planes or airports. In addition … were the horrific bombings of Pan Am 103 and UTA 772, the bombing of an Air India 747 over the North Atlantic that killed 329 people, and the saga of TWA Flight 847.

We’ve been dealing with issues of air terrorism for years now, but why in the world have we gone so crazy insane about it? There are over 50,000 traffic fatalities in the US every year, but we accept that in exchange for the ability to drive at 70 mph. Instead of theatre there are things that can be done to minimize the risk to air travelers without treating them like baggage.

Second, there is the plane flight itself. Gone are complimentary snacks and in some cases complimentary beverages. The planes are cramped and often dirty, as there is no time between flights to clean them properly.

Finally, financial conditions are such that many planes in the air are rather old. I know pilots and understand from them that even old planes that are well maintained are very safe, but I’ve personally experienced problems with the aging American Airlines fleet of MD-80s which results in delays and cancellations as repair parts have to be brought in (the problem being dutifully caught in the pre-flight check).

This subject is on my mind since tomorrow I fly for the first time in 2010. It will be a little less traumatic for me since my elite status gets me upgraded to first class. Even then, I still was nearly molested by an American Airlines flight attendant and have never received an acknowledgement to the letters I wrote complaining about it.

What’s the use of working toward elite status if you don’t want to fly?

Luckily for me, OpenNMS has grown to the point that I don’t have to be on the road so much. We have others who can do what I used to do, and in fact do it better.

Deep down I still view air travel as something magical. I mean, seriously, tomorrow I get on a plane at 9am and I’m in LA by 3pm, local time. There I get to meet up with Gareth, Ilan and the rest of the SCaLE gang, get to attend one of my favorite conferences of the year, and I hear the keynote on Sunday is just awesome.

I get to sit in a chair in the sky.

It’s even more amazing when I travel to another country. Get on a plane, take a nap, and suddenly I’m elsewhere, with a new language to learn, new friends to make and new things to do.

But my experiences and those of Kevin Smith have me checking out Amtrak train schedules, and it is doubtful I’ll be on Southwest any time soon.

I have enough problems without being “too fat to fly”.

On Plugins

This morning we were trying to set up a GoToMeeting so that Jay could demo the new JasperReports integration into OpenNMS (for those of you playing at home it should be merged into trunk this afternoon).

We couldn’t authenticate.

The http://status.gotomeeting.com site displayed the following:

Whenever I have an issue with a network resource, my first reaction is to get OpenNMS to monitor it.

With many monitoring tools, there is a lot of talk about “plugins”. In yesterday’s post about Nagios, one of the reasons it is so popular is that it is really easy to add custom scripts.

But there is a downside. Since it is so easy to add code, it is much harder to insure that the plugins both perform well with respect to scaling and that they are secure.

This is one of the reasons we build in very generic monitors and data collectors into OpenNMS versus having an open plugin architecture (for example, recently Slashdot had an article claiming that “insecure plugins are a serious threat to all browsers“). The idea is that you should be able to do much of what you need simply by configuring, or at the most modifying, an existing monitor, and those monitors are designed to scale.

I remember at one client they were using Nagios and they had a check script that made a number of SNMP queries to determine the status of RAID controllers on Dell servers. I ported the script to OpenNMS using the general purpose monitor, we pointed it at the entire network (previously, they had only used it on a small number of servers) and went to lunch.

When we came back, the system had automatically discovered over 1400 servers and the load was around 20 (on a pretty powerful box). This wouldn’t do, so we rewrote the script by modifying the existing SNMP monitor to duplicate this functionality (these became the PERC and OMSA monitors). After we switched, the load on the system was less than 1. All of those fork-exec calls were killing the system.

Well, back to GoToMeeting. Once their system was back to normal, the status page looked something like this:

I decided that it would be pretty easy to use the Page Sequence Monitor to test for the string “The GoToMeeting service is currently available”. If that string was on the page, I will assume the service is up.

First, I added the service to the capsd-configuration.xml file:

<protocol-plugin protocol="Go2Status"
     class-name="org.opennms.netmgt.capsd.plugins.LoopPlugin"
     scan="off" user-defined="false" />

I used the LoopBack plugin since I had decided to manually add the service to the GoToMeeting node, so I had no need for automatic discovery.

Then I added it to the poller-configuration.xml file:

<service name="Go2Status" interval="300000" user-defined="true" status="on">
    <parameter key="retry" value="1"/>
    <parameter key="timeout" value="5000"/>
    <parameter key="page-sequence">
      <page-sequence>
        <page path="/"
            port="80"
            virtual-host="status.gotomeeting.com"
            successMatch=
                 "(?s).*The GoToMeeting service is currently available.*" />
      </page-sequence>
    </parameter>
</service>

(don’t forget to add the monitor tag at the bottom)

<monitor service="Go2Status"
     class-name="org.opennms.netmgt.poller.monitors.PageSequenceMonitor"/>

And finally, I used the provisioner to add it to the IP address for “status.gotomeeting.com”:

Voila! I now have a monitor that I know is secure and will perform well and I was able to do this strictly through configuration.

By leveraging the existing monitors and collectors in OpenNMS, sometimes augmented by agent technology such as Net-SNMP, it is quite easy to extend OpenNMS without the need to introduce variables that could degrade either security (such as using ssh and shared keys to access remote hosts) or performance (such as running scripts that require a shell).