Do you know why the 3 nodes are down ? With replication, the copy of data that were hosted on those failed nodes will not be available. However, the data will still be served by the hosts having the other 2 copies - so I don't think you need to copy the data again.
Unless for some reason the 3 copies of some data ended up on these nodes, in which case those will not be available Maybe you could do a ' hadoop fsck /' to confirm if the HDFS is healthy. -Rajesh On Wed, Feb 25, 2015 at 9:21 AM, tesm...@gmail.com <tesm...@gmail.com> wrote: > Dear all, > > I have transferred the data from local storage to HDFS in my 10 nodes > Hadoop cluster. The relication facotr is 3. > > Some nodes, say 3, are not available after some time. I can't use those > nodes for computation or storage of data. > > What will happen to the data stored on HDFS of those nodes? > > Do I need to remvoe all the data from HDFS and copy it again? > > Regards, > >
