Our Community is getting an upgrade! To get everything ready for the relaunch, we’ll be placing the site in read-only mode starting September 21st.
We really appreciate your understanding while we get things set up behind the scenes. Catch up on all the exciting details about the move here.
Need help or have questions? Drop us a line at [email protected]

Archives of Support Questions (Read Only)

This is an archived board for historical reference. Information and links may no longer be available or relevant
Announcements
This board is archived and read-only for historical reference. To ask a new question, please post a new topic on the appropriate active board.

Hbase region servers load balancing in a cluster with disparate nodes

avatar
Explorer

In our CDH 5.5 cluster, we have about 12 region servers nodes out of which 4 nodes are 8 cores and the remaining are 16 cores. We have noticed those 4 nodes are always busy during our load. Is it possible to assign weights to these low powered 4 nodes such that they will house let say 50% less regions compared to other set of nodes.

1 ACCEPTED SOLUTION

avatar
Master Collaborator

Short Answer: No, not possible to balance with respect to cpu/memory resources out of the box today to my knowledge.

 

Longer Answer: You can write a custom balancer either external to hbase or using the hbase balancer protocol. Using the region_mover.rb as an example you can write your own jruby that can be run by the shell. Unfortunately, ultimately you will likely be better off without the underpowered nodes in the cluster than you are with them in.  Perhaps keep them in for HDFS storage and run just YARN there?

View solution in original post

1 REPLY 1

avatar
Master Collaborator

Short Answer: No, not possible to balance with respect to cpu/memory resources out of the box today to my knowledge.

 

Longer Answer: You can write a custom balancer either external to hbase or using the hbase balancer protocol. Using the region_mover.rb as an example you can write your own jruby that can be run by the shell. Unfortunately, ultimately you will likely be better off without the underpowered nodes in the cluster than you are with them in.  Perhaps keep them in for HDFS storage and run just YARN there?