-Re: hadoop 2.2.0 nodemanager graceful stop
Harsh J 2014-05-19, 09:26
There is no way to do a graceful stop presently. You can follow
https://issues.apache.org/jira/browse/YARN-914 for tracking this feature
That said, a simple restart that does cause container failures is still OK
with most apps (especially with MR) since these apps have failure tolerance.
On Mon, May 19, 2014 at 7:58 AM, Shengjun Xin <[EMAIL PROTECTED]> wrote: