[
https://issues.apache.org/jira/browse/HBASE-26260?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Andrew Kyle Purtell updated HBASE-26260:
----------------------------------------
Description:
Once we are guaranteed that a LZ4 implementation will be universally available,
we can switch the very pessimistic and essentially unusable default of GZ codec
for WAL value compression (HBASE-25869), chosen because GZ was formerly the
only universally available option, to LZ4, which is very fast.
Originally the idea was to use SNAPPY, but LZ4 (in hbase-compression-lz4)
outperforms all Snappy alternatives in microbenchmarking.
was:
Once we are guaranteed that a SNAPPY implementation will be universally
available, we can switch the very pessimistic and essentially unusable default
of GZ codec for WAL value compression (HBASE-25869), chosen because GZ was
formerly the only universally available option, to SNAPPY, which is the
strongly preferred choice for that feature.
> Update default compression codec for WAL value compression to LZ4 (via
> lz4-java)
> --------------------------------------------------------------------------------
>
> Key: HBASE-26260
> URL: https://issues.apache.org/jira/browse/HBASE-26260
> Project: HBase
> Issue Type: Sub-task
> Components: wal
> Reporter: Andrew Kyle Purtell
> Assignee: Andrew Kyle Purtell
> Priority: Major
> Fix For: 2.5.0, 3.0.0-alpha-2
>
>
> Once we are guaranteed that a LZ4 implementation will be universally
> available, we can switch the very pessimistic and essentially unusable
> default of GZ codec for WAL value compression (HBASE-25869), chosen because
> GZ was formerly the only universally available option, to LZ4, which is very
> fast.
> Originally the idea was to use SNAPPY, but LZ4 (in hbase-compression-lz4)
> outperforms all Snappy alternatives in microbenchmarking.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
