On 28/11/11 08:01, mohmmadanis moulavi wrote:
> Where will i get Developers Documentation of hadoop.
> provide me some link.
it's the source tree, with some more on http://wiki.apache.org/hadoop
There are some big assumptions in the code that you already know
networking, Java, testing, etc -if that isn't the case, this codebase is
not the place to learn.
Similarly, everyone is nervous about changes to the bits of the system
that are considered critical: the file system, the public apis, the job
If you do want to get into the codebase
-start with your own applications, follow through how they work in the
in-JVM mini clusters, see how code is executed, data is read, etc.
-look at the IPC infrastructure, serialization, and other things that
are foundational. Even if you don't go near this code, you will see it
often enough that you need to understand it.
-if you want to do things like new scheduling, block placement
algorithms, there are plug in points for these -so you can do stuff here
without risk to anyone's working systems other than your own.
-Otherwise: start with the small bugs that are irritating you. Even text
messages, things there aren't enough tests for, diagnostics, etc. More
diagnostics are things everyone needs