[
https://issues.apache.org/jira/browse/HBASE-17869?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15955986#comment-15955986
]
Jerry He commented on HBASE-17869:
----------------------------------
Yeah, I had the same question. Two code paths (true or false) should yield the
same results.
Since the 'true' route is more commonly used, can we assume it has the correct
results?
Then there could be a bug somewhere in the 'false' route. In the FuzzyRowFilter?
{code}
java.lang.AssertionError: expected:<YES> but was:<NO_NEXT>
at org.junit.Assert.fail(Assert.java:88)
at org.junit.Assert.failNotEquals(Assert.java:834)
at org.junit.Assert.assertEquals(Assert.java:118)
at org.junit.Assert.assertEquals(Assert.java:144)
at
org.apache.hadoop.hbase.filter.TestFuzzyRowFilter.testSatisfiesForward(TestFuzzyRowFilter.java:80)
{code}
{code}
java.lang.AssertionError: expected:<YES> but was:<NEXT_EXISTS>
at org.junit.Assert.fail(Assert.java:88)
at org.junit.Assert.failNotEquals(Assert.java:834)
at org.junit.Assert.assertEquals(Assert.java:118)
at org.junit.Assert.assertEquals(Assert.java:144)
at
org.apache.hadoop.hbase.filter.TestFuzzyRowFilter.testSatisfiesReverse(TestFuzzyRowFilter.java:120)
{code}
> UnsafeAvailChecker wrongly returns false on ppc
> -----------------------------------------------
>
> Key: HBASE-17869
> URL: https://issues.apache.org/jira/browse/HBASE-17869
> Project: HBase
> Issue Type: Bug
> Affects Versions: 1.2.4
> Reporter: Jerry He
> Assignee: Jerry He
> Priority: Minor
> Fix For: 2.0.0, 1.4.0
>
> Attachments: HBASE-17869.patch
>
>
> On ppc64 arch, java.nio.Bits.unaligned() wrongly returns false due to a JDK
> bug.
> https://bugs.openjdk.java.net/browse/JDK-8165231
> This causes some problem for HBase. i.e. FuzzyRowFilter test fails.
> Fix it by providing a hard-code workaround for the JDK bug.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
