Re: Contributing

Lars Marowsky-Bree <[email protected]> Thu, 10 Oct 2002 17:05:11 +0200
Newsgroups gmane.linux.failsafe
Message-ID <[email protected]>
On 2002-10-10T11:55:30,
   Dhammika Pathirana <[email protected]> said:

> I'm looking to get some experience in cluster programming and would like 
> to get involved. Though I'm quite conversant in linux programming I 
> don't have any previous experience in clustering. I'd like to get some 
> advice on possible starting points and possible individual projects.

Hi Dhammika, this is good news. (Able people willing to help always are!)

As you might have read, I personally don't consider Linux FailSafe the best
place to apply much effort right now.

_However_, there is a "Open Clustering Framework" effort going on, where we
strive to a) define a common API for all things clustering on Linux
(comparable to the Linux Standard Base), and b) also provide a reference
implementation, based on heartbeat.

To do this, we have identified a rather generic "architecture" of components
common to most clustering solutions and are now working through them. ie,
discussing a sensible API and also coding the reference code base.

(The later not always being coded "up from scratch", but by reusing existing
code and adapting it; ie, group communications code could well rescue some
algorithms from FailSafe or DLM be based upon the Distributed Lock Manager
open sourced by IBM)

See http://lars.marowsky-bree.de/dl/ocf.pdf for a toplevel paper I wrote on
the OCF.

So, this would be the perfect place to go for a "bite sized" chunk of
clustering work ;-)


Sincerely,
    Lars Marowsky-Brée <[email protected]>

-- 
Principal Squirrel
Research and Development, SuSE Linux AG
 
``Immortality is an adequate definition of high availability for me.''
	--- Gregory F. Pfister