On Wed, Sep 4, 2013 at 8:41 PM, Fabrice Bacchella
<fbacche...@spamcop.net> wrote:
> Hi, I'm the main developper on rrd4j.
>
> I've been very busy lately on this project, and talked with Tarus Balog about
> licensing issue.
>
> I'm happy that you seems to be open about an Apache License, and look forward
> for more exchange between our projets.
>
> I have already a lot of improvement and bug fix in rrd4j, with some heavy
> perfomance one by changing the disk layout of rrd files. The pool is more
> scalable I think.
>
> For now, I have 3 open projects on rrd4j, and wanted to inform you about
> them, I hope it can help you improving jrobin and opennms.
>
> - The first one is an upgraded of the API to be more on sync with rrdtool,
> and it made a confusion in the API between consolidation function and VDEF, I
> hope I improved the situation, it can also provide access to custom VDEF.
> It's currently in the branch master of rrd4j and when it will be enought
> tested I plan to release it as rrd4j 2.3
> - The backend API was not very rich, it provides only raw byte storage. I'm
> trying to write something more high level, that should allow me to write a
> cassandra back end wich is a very important goal. It should also be easier to
> add custom backends as plugin. It's in the NewBackEnd branch. When it will be
> finished and tested, I plan to release it as rrd4j 3.
> - The JRobinRrdStrategy strategy expose only a very small subset of what
> rrdtool can provide, that's a pity and don't help increase jrobin/rrd4j
> usage, writing a RRD4JRrdStrategy is not a difficult task, just tedious, I
> have started it.
>
> I have also written a caching backend for rrd4j, that can use directio on
> Solaris/Linux plateform, for high performance collect.
>
> For all of this, if any of you is interested, help, comments and ideas are
> welcome, and I'm open to any exchange you might be intersted in.
The piece I've always found missing is a tool to build aggregate
totals that track load balanced and failover sets. A simple example
would be a stacked graph of the bandwidth of several internet-facing
interfaces on different routers, but it gets messier when you want
some metric summarized across a cloud of servers where the individual
nodes change frequently (so the opennms symlink trick doesn't work)
and you want to track a continuous history of a total as well as each
individual node. And you want it to track at each sample interval so
as not to double up on failovers.. This problem may not relate well
to the low-level library that just stores and retrieves values but it
must be a common problem these days. Do you know any way to address
it? In the context of opennms it would be ideal if the data
collection side could watch the aggregate value during the input so
you could have event thresholding at that level but even being able to
store/view a history would be useful.
--
Les Mikesell
lesmikes...@gmail.com
------------------------------------------------------------------------------
Learn the latest--Visual Studio 2012, SharePoint 2013, SQL 2012, more!
Discover the easy way to master current and previous Microsoft technologies
and advance your career. Get an incredible 1,500+ hours of step-by-step
tutorial videos with LearnDevNow. Subscribe today and save!
http://pubads.g.doubleclick.net/gampad/clk?id=58041391&iu=/4140/ostg.clktrk
_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ
opennms-devel mailing list
To *unsubscribe* or change your subscription options, see the bottom of this
page:
https://lists.sourceforge.net/lists/listinfo/opennms-devel
