Just one thing. Let's do it as Doug proposed earlier
hadoop dfs -check
This is a good piece of code, but too far from what the real
fsck should do to deserve the name :-)
I mean sooner or later somebody will write a real fsck, and we will have
to call it something else.
--Konst
Andrzej Bialecki wrote:
Doug Cutting wrote:
Andrzej Bialecki wrote:
I would like to commit DFSck before the release, if that's ok.
That sounds fine to me.
This week, pre-release, we should be careful: (1) not to make changes
that might destabilize things; and (2) we should not add new public
APIs that have not been well-vetted, since once these are in a
released version it will be harder to modify them. But DFSck is
primarily a new class that prints something informative, right? So
it shouldn't break anything, and it doesn't add any public API other
than DFSck().main(), right? (I can't get through to Jira right now...)
Right, it doesn't touch anything in the API, just adds a shortcut
"hadoop fsck" to bin/hadoop.
