hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Hudson (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-10212) New rpc metric: number of active handler
Date Tue, 20 May 2014 14:18:42 GMT

    [ https://issues.apache.org/jira/browse/HBASE-10212?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14003348#comment-14003348

Hudson commented on HBASE-10212:

FAILURE: Integrated in hbase-0.96 #399 (See [https://builds.apache.org/job/hbase-0.96/399/])
HBASE-10561 Forward port: HBASE-10212 New rpc metric: number of active handler (liangxie:
rev 1594131)
* /hbase/branches/0.96/hbase-hadoop-compat/src/main/java/org/apache/hadoop/hbase/ipc/MetricsHBaseServerSource.java
* /hbase/branches/0.96/hbase-hadoop-compat/src/main/java/org/apache/hadoop/hbase/ipc/MetricsHBaseServerWrapper.java
* /hbase/branches/0.96/hbase-hadoop1-compat/src/main/java/org/apache/hadoop/hbase/ipc/MetricsHBaseServerSourceImpl.java
* /hbase/branches/0.96/hbase-hadoop2-compat/src/main/java/org/apache/hadoop/hbase/ipc/MetricsHBaseServerSourceImpl.java
* /hbase/branches/0.96/hbase-server/src/main/java/org/apache/hadoop/hbase/ipc/MetricsHBaseServerWrapperImpl.java
* /hbase/branches/0.96/hbase-server/src/main/java/org/apache/hadoop/hbase/ipc/RpcServer.java
* /hbase/branches/0.96/hbase-server/src/test/java/org/apache/hadoop/hbase/ipc/MetricsHBaseServerWrapperStub.java
* /hbase/branches/0.96/hbase-server/src/test/java/org/apache/hadoop/hbase/ipc/TestRpcMetrics.java

> New rpc metric: number of active handler
> ----------------------------------------
>                 Key: HBASE-10212
>                 URL: https://issues.apache.org/jira/browse/HBASE-10212
>             Project: HBase
>          Issue Type: Improvement
>          Components: IPC/RPC
>            Reporter: Chao Shi
>            Assignee: Chao Shi
>             Fix For: 0.94.17
>         Attachments: hbase-10212.patch
> The attached patch adds a new metric: number of active handler threads. We found this
is a good metric to measure how busy of a server. If this number is too high (compared to
the total number of handlers), the server has risks in getting call queue full.
> We used to monitor  # reads or # writes. However we found this often produce false alerts,
because a read touching HDFS will produce much high workload than a block-cached read.
> The attached patch is based on our internal 0.94 branch, but I think it pretty easy to
port to rebase to other branches if you think it is useful.

This message was sent by Atlassian JIRA

View raw message