Thanks, I have done this by copying meta data from one local directory to another on each node, an interesting thing is I need to add "-upgrade" command option when starting NN.
On Wed, Feb 19, 2014 at 10:10 AM, sudhakara st <[email protected]>wrote: > Hello Anfernee, > > Actually this not data migrations !! . It is upgrade. If you have to > cluster older one 0.21.0 with data and another one 2.2.0 is new. Then if > want to copy data from older to newer then it is data migration, i doesn't > require any extra care just copy data. if want change in the same old > cluster then you have to follow upgrade steps > please refer these: > http://wiki.apache.org/hadoop/Hadoop_Upgrade > The Hadoop definitive guide - Tom white 3rd edition page -360 > http://hortonworks.com/blog/running-existing-applications-on-hadoop-2-yarn/ > > > > > On Wed, Feb 19, 2014 at 4:30 AM, Anfernee Xu <[email protected]>wrote: > >> Hi, >> >> I have an existing Hadoop cluster built from 0.21.0 release and have some >> important data on HDFS, now I'm planning to move to the current GA >> release(2.2.0), what's the procedure for the data migration(from 0.21.0 >> hdfs to 2,20 hdfs) if topology is not changed(all nodes are not changed)? >> Can I simply copy all meta data files from 0.21.0(say >> /hadoop_021/temp/hdfs/namenode/) to 2.2.0(say >> /hadoop-220/temp/hdfs/namenode). >> >> Thanks >> >> -- >> --Anfernee >> > > > > -- > > Regards, > ...sudhakara > > -- --Anfernee
