hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Appy (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-15673) [PE tool] Fix latency metrics for multiGet
Date Tue, 19 Apr 2016 03:08:25 GMT

    [ https://issues.apache.org/jira/browse/HBASE-15673?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15247069#comment-15247069
] 

Appy commented on HBASE-15673:
------------------------------

Ready for commit.
The failing test is flaky. From the output:
{noformat}
Flaked tests: 
org.apache.hadoop.hbase.regionserver.wal.TestAsyncLogRolling.testLogRollOnDatanodeDeath(org.apache.hadoop.hbase.regionserver.wal.TestAsyncLogRolling)
  Run 1: TestAsyncLogRolling.testLogRollOnDatanodeDeath:63 expected:<1> but was:<0>
  Run 2: PASS
{noformat}

> [PE tool] Fix latency metrics for multiGet
> ------------------------------------------
>
>                 Key: HBASE-15673
>                 URL: https://issues.apache.org/jira/browse/HBASE-15673
>             Project: HBase
>          Issue Type: Bug
>            Reporter: Appy
>            Assignee: Appy
>         Attachments: HBASE-15673-master.patch
>
>
> we write out latency value after each row is processes by testRow() function. But if
multiget is enabled, say set to 10, testRow() returns immediately for 9 rows and sends out
request in the 10th iteration. This screws up latency metrics (50th, 75th, 90th percentiles
are all 0).



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message