[rtg] rated, new version of rtgpoll

[Jacob Steinberger]

Why have it as a fork? Why not just update the new RTG code base that  
was made?

Since someone is trying to manage the product, I'd hate to see forks  
when we could keep it all 'in house'.

Jacob

Quoting Matt Provost <mprovost at termcap.net>:

> So I've just published a new project called 'rated' which is a fork of
> RTG, specifically rtgpoll.
>
> http://github.com/mprovost/rated/tree/master
>
> Basically I've pulled out all of the supporting scripts and graphing
> code. I've made a bunch of changes that I've had in mind for years for
> RTG.
>
> Some of the improvements:
>
> * Fixed several memory leaks.
> * Fixed autotools linking so the database libraries only link against
> the correct shared libraries. Before everything used a single list so
> even rtgpoll was linking against the mysql and postgres libraries. This
> really reduces the memory footprint of the daemon.
> * General code cleanup - took out a lot of IFDEFs, spacing, etc.
> * Signal handling is much cleaner.
> * Database disconnection - it should now handle having the database
> disconnect and reconnect when it can. For counters it also just keeps
> incrementing the counter in memory, so you shouldn't lose any data (just
> precision) while the db is unavailable.
> * Lots of cleanup in the snmp polling loop.
> * Better debug output.
>
> It's been mostly tested against postgres but mysql should also work
> fine. I haven't tested the oracle driver but that's on my list. I'm sure
> there are still lots of references to rtg that I haven't cleaned up yet.
>
> This first release hasn't changed the codebase very much. It should
> still be a drop-in replacement for rtgpoll. In the future I've got some
> ideas for this that will take it in other directions. I'd like to get
> the targets moved into a database. I'm also interested in a compressed
> backend. I've begun playing with a scheduler that can vary the polling
> interval dynamically based on the rate of change of the data. I'm not
> sure if everyone will like the changes which is why I decided to make a
> fresh start.
>
> I've seen RTG stagnate in the past so hopefully this will be a more
> public development process. If anyone has any ideas please fork the code
> on github!
>
> Thanks,
> Matt
> _______________________________________________
> RTG mailing list
> RTG at lists.grdata.com
> http://lists.grdata.com/mailman/listinfo/rtg
>