[GORA-413] Support creation of dynamic columns within Gora datastore mapping designs

• Type: New Feature
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 0.6
• Fix Version/s: 1.0
• Component/s: gora-hbase
• Assignee: Unassigned
• Reporter: Lewis John McGibbney
• Created: 26/Feb/15 20:05
• Updated: 03/Mar/22 14:27

Description

The conversation taking place on dynamically generating HBase columns has raised an issue that new functionality needs to be added in order to achieve this.
The main driver for this issue coming to light is that Chukwa logs need to dynamically create many many columns over time directly dependent on the number of data chunks we get. Each data chunk has a [Sequence ID], this sequenceID should be the column name.

The table design will look like this

Row Key: [Invert Date]:[Data Type]:[Primary Key]
Column Family: log
Column Name: [Sequence ID]
Timestamp: [log entry timestamp]

Example:

Row Key: 2132013102:TT:host1.example.com
Column Family: log
Column Name: 1230
Cell Value: 2013-01-23 12:01:30 INFO This is a log entry.
Timestamp: 1358942490

The inverted date allow the table to be partitioned by hour or day of the month or month more easily.
The usage of column name for consecutive sequence to allow fast retrieval in a linear scan. This format is typically good for retrieve a hour worth of logs fast for a node. Hence, if we are doing batch scanning of the table in a rolling window via map reduce job at every hour interval, we get a even spread the work load to multiple map reduce tasks.