[
https://issues.apache.org/jira/browse/HBASE-14188?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14659588#comment-14659588
]
ramkrishna.s.vasudevan commented on HBASE-14188:
------------------------------------------------
The test failure is due to
{code}
java.lang.OutOfMemoryError: unable to create new native thread
at java.lang.Thread.start0(Native Method)
at java.lang.Thread.start(Thread.java:713)
at
org.apache.hadoop.hbase.MultithreadedTestUtil$TestContext.startThreads(MultithreadedTestUtil.java:61)
at
org.apache.hadoop.hbase.io.hfile.CacheTestUtils.hammerSingleKey(CacheTestUtils.java:200)
at
org.apache.hadoop.hbase.io.hfile.bucket.TestBucketCache.testCacheMultiThreadedSingleKey(TestBucketCache.java:177)
{code}
Seem unrelated. It is passing locally.
bq.getNextCellStartPosition -> Now it is not returning the next cell's start
pos. It is returning cur cell serialization size in the HFile.. Pls change the
method name accordingly. Can fix on commit.
Changed this to getCurCellSize().
> Read path optimizations after HBASE-11425 profiling
> ---------------------------------------------------
>
> Key: HBASE-14188
> URL: https://issues.apache.org/jira/browse/HBASE-14188
> Project: HBase
> Issue Type: Sub-task
> Components: Scanners
> Reporter: ramkrishna.s.vasudevan
> Assignee: ramkrishna.s.vasudevan
> Attachments: HBASE-14188.patch, setSeqId.png
>
>
> This subtask deals with some improvments that can be done in the read path
> (scans) after the changes for HBASE-11425 went in.
> -> Avoid CellUtil.setSequenceId in hot path.
> -> Use BBUtils in the MultibyteBuff.
> -> Use ByteBuff.skip() API in HFileReader rather than
> MultiByteBuff.position().
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
