Following the recent project split, the bin folder is still in the
common project. While it's possible to start the namenode and
datanode(s) directly from Eclipse, the general workaround is to just
copy the bin folder from a checkout of the common/core project into the
hdfs folder and run compiled jar from there. This will be fixed at some
point in the future.
Hadoop at Yahoo!
Thanh Do wrote:
> I have problem with running hdfs. Basically, I was able to check out
> hdfs/chunk and build it successfully in eclipse. But there was no bin
> directory containing useful scripts (like in the tarball release) so
> that I can use to run my modified version of hdfs. Any suggestion is