What is meant by Hot spotting in Hbase?
The most common cause for hotspotting is inserting rows with monotonically increasing row keys.in other words It happens most of the time due to sequential row key design, in hbase set of row resides on region servers and when we have a sequential row keys it tries to write the rows on a single region server, which result in huge amount of write to a single and result in a kind of congestion and region server to be overloaded. So we design a row key in such a way that it is distributed in a better way through out the region servers.
In that case only the last region will get the writes and no amount of splitting will fix that (only one region serer will hold the last region of the table regardless of how small it is).
The HBase record table is where you persist the record data. While creating the record table, there are several elements that can impact performance.
The considerations are:
If any of these elements are not set properly, you might see poor performance when loading the data. You might also see hotspotting (hotspotting is when only one of the nodes uses high-CPU utilization despite having multiple nodes in the cluster). Hotspotting can occur simply when loading data or even when running Big Matchbatch jobs like Derive, Compare, or Link.
Retrieving data ...