Programming Assignment 3

Click here to view full description of the assignment.


Hbase not starting shown zookeeper instances running on nodes.

In this case, you will need to log into every regionserver node, as well as hbase master node and type "jps" in the terminal to see any Hbase related process running on that node like (HQuorunPeer/HRegionServer etc). Note their process id and kill them using the command kill -9 PROCESSID.

Refactor script

I have uploaded a refactoring script to get rid of any temporary log files that gets generated from a botched start of hbase. In case you try to update your configurations and restart hbase, it is better to run this script before doing so.
Download the script here. Copy it into the conf directory of hbase. Please change YOURUSERNAME string with your actual username in the .sh file.
Before running the script, please make sure that the "test_home" variable inside this file points to the conf directory of your hbase installation. Create a file called "allmachines" inside your conf that contains the names of all your regionservers and your hbase master that you have mentioned in the hbase-site.xml. The way to run this is similar to how to did for hadoop(make sure you are running from the terminal of the machine you are on).

HBase not starting- log showing issues with leaderport at 3888.

In this case, you will need to add a new property called hbase.zookeeper.leaderport in your hbase-site.xml file and then provide a unique port in the portrange assigned to you in assignment 0. Now, to be sure, run the stop hbase command from terminal. Also, you will need to run the refactor script mentioned above like you did for hadoop. You will now need to restart HBase. It is better to stop everything and then start from starting hadoop, yarn and then hbase. Also, make sure that no Hbase related processes are still running on your regionservers.

Page last modified on April 11, 2017, at 08:50 AM EST