[
https://issues.apache.org/jira/browse/HBASE-18294?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16068102#comment-16068102
]
Eshcar Hillel commented on HBASE-18294:
---------------------------------------
While benchmarking, we noticed the number of flushes in stores with basic
compaction is similar to the number of flushes in stores with no compaction. We
expected the number of flushes with basic compaction to be smaller as the index
is smaller and therefore the heap size is smaller and the memstore can absorb
more data before it is required to flush to disk.
This is a very small change, I will make a patch later today.
However, it might have significant affect on performance, and therefore should
be at least considered also for 2.0.0.
> Flush policy checks data size instead of heap size
> --------------------------------------------------
>
> Key: HBASE-18294
> URL: https://issues.apache.org/jira/browse/HBASE-18294
> Project: HBase
> Issue Type: Bug
> Reporter: Eshcar Hillel
> Assignee: Eshcar Hillel
>
> A flush policy decides whether to flush a store by comparing the size of the
> store to a threshold (that can be configured with
> hbase.hregion.percolumnfamilyflush.size.lower.bound).
> Currently the implementation compares the data size (key-value only) to the
> threshold where it should compare the heap size (which includes index size,
> and metadata).
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
