The snappy integration in Hbase is based on Hadoop-Snappy (
Hadoop trunk has now a Snappy codec (based on Hadoop-Snappy).
Until we have a release of Hadoop including the Snappy codec, to enable
Hbase to use Snappy, Hbase should continue using Hadoop-Snappy.
To ensure that things work smoothly and the switch over does not create any
issues, Hadoop-Snappy should be kept in sync with the Snappy codec in Hadoop
With the help of Issay, we've synch Hadoop-Snappy with Hadoop trunk.
I'll take the job to kept them in sync in sync until we have a Hadoop
release with the Snappy codec.