[
https://issues.apache.org/jira/browse/HBASE-17575?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15848371#comment-15848371
]
Eshcar Hillel commented on HBASE-17575:
---------------------------------------
The bug fix in this patch is in MemStoreSegmentsIterator.
The getScanners method is changed to initiate segment scanners with
readPoint=Integer.MAX_VALUE instead of store.getSmallestReadPoint().
We initially thought that the value returned by store.getSmallestReadPoint
method is the MVCC read point but the tests revealed that the actual value
returned is smaller than the read point so the scanner may end up filtering out
versioned that should be returned to the caller. Using Integer.MAX_VALUE fixes
this problem, and is similar to what is used in case of flush to disk.
> Towards Making BASIC the Default In-Memory Compaction Policy
> ------------------------------------------------------------
>
> Key: HBASE-17575
> URL: https://issues.apache.org/jira/browse/HBASE-17575
> Project: HBase
> Issue Type: Sub-task
> Reporter: Eshcar Hillel
> Assignee: Eshcar Hillel
> Attachments: HBASE-17575-V01.patch
>
>
> In preparation for switching the default in-memory compaction policy to
> BASIC, we would like to make sure that once we do the switch all tests will
> run (and pass!) with the new default policy.
> To this end, we remove the NONE configuration setting added to tests in
> HBASE-17294 and HBASE-17316.
> We verified these tests pass with all 3 memory compaction policies.
> For each test
> (1) if all 3 policies pass the test -- we remove the configuration from the
> test.
> (2) if some fail we add tests of all 3 configurations, e.g., by parameterized
> tests. When needed we update expected results.
> One test failure identified a small bug which is also fixed in the patch.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
