Is there anything cooler or more badass than writing software used in space? Whenever I'm told that a piece of software isn't fit for "mission critical applications," I think of NASA.
I'd love to help out with this, but I'm afraid I lack the necessary skill set. Still, maybe I will poke around the NDOUtils source anyways.
NASA's commitment to open source is fairly great. And, postgresql is a really great database. If you haven't used it before, I highly recommend giving it a run.
Well, as far as I can tell, all the other solutions they could be using suck more, at least adding new modules to nagios is about as easy as writing a script. Cause that's all you have to do. I've dealt with installs checking hundreds of servers,routers and switches and our only major issue was needing to distribute the checks onto multiple machines. One box will die trying to ping,connect,... onto too many nagios boxes.
My guess would be that NASA already uses postgres for other duties and is hesitant to introduce a new database into their infrastructure - especially one with so many known problems.
If I was them then I most certainly wouldn't want a database with silent data truncation/corruption issues to come anywhere near my spaceships. Not even for seemingly unimportant tasks because even those may cost amazing amounts of money to fix if they go wrong a few thousand kilometers above groundlevel...
PostgreSQL is well-known for it's focus on data integrity and durability first, performance second. The team didn't really focus heavily on performance until the 8.x release series, after the database had been proven as extremely stable. Also, it has a much more complete implementation of SQL and larger feature set.
I've used both very recently, and I would definitely avoid MySQL given a choice.
From my experience, postgres is faster, especially at more difficult problems (complex joins). I've never had issues in postgres joining several tables into 10 million plus row relations, and then filtering down to the set of interest.
I did a recent test adding a similar column to a MySQL table with 3 million rows in it compared to adding a column to a postgres table with 30 million rows in it. MySQL took 3000 seconds, and postgres took 82 ms.
Functional indexes. Doing things like "create unique index user_emails on users (lower(email));" works great in postgres.
Isn't the computing power on the ISS almost purely made up of Thinkpad laptops running Windows? Will they be running PostgreSQL and Nagios on a Windows machine?
That's awesome. Nagios in space. Sounds like a cool comedy title. Were about to install Nagios here too. One more point for management to like the idea ;p
I'm an intern at NASA this summer, and I've actually convinced the team I'm working with to use Redis/Resque for the application we're working on. Unfortunately it's 'just' a web application ;)
"I felt about as good as anybody would, sitting in a capsule on top of a rocket that were both built by the lowest bidder."
-- John Glenn on how it felt waiting to for the rocket to launch
