Yeah, just to reinforce what Ted is saying, DO NOT run HDFS's balancer if you use HBase. Doing so will move blocks in such a way as to destroy data locality and negatively impact HBase performance (until a major compaction in HBase is done).
On Friday, October 7, 2016, Ted Yu <yuzhih...@gmail.com> wrote: > For #1, it depends on whether major compaction is disabled. If major > compaction is enabled, timing of major compaction would affect the data > locality. > > For #2, no. hdfs rebalance is orthogonal to hbase. > > For #3, perform major compaction at earliest convenience. > > On Thu, Oct 6, 2016 at 11:47 PM, Manjeet Singh <manjeet.chand...@gmail.com > <javascript:;>> > wrote: > > > Hi All, > > I have question on re balance, my query is how hbase rebalancing the data > > after adding new nodes in cluster > > <http://stackoverflow.com/questions/23686387/hadoop-and- > > hbase-rebalancing-after-node-additions> > > ? > > > > > > > > 1. Do I need to explicitly rebalance hbase after adding the new node > in > > cluster? > > > > 2. On my cloudera I have hdfs rebalance does its take care of hbase > data > > to be balance? > > > > 3. What is the best way to make sure that both hadoop and hbase are > > rebalanced and work fine? > > > > > > Thanks > > > > Manjeet > > > > -- > > luv all > > > -- -Dima