Mailing-List: contact issues-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Date: Tue, 21 Aug 2012 12:50:38 +1100 (NCT)
From: "Jonathan Leavitt (JIRA)" <jira@apache.org>
To: issues@hbase.apache.org
Message-ID: <138331892.33117.1345513838153.JavaMail.jiratomcat@arcas>
In-Reply-To: <1788723162.1690.1344378070529.JavaMail.jiratomcat@issues-vm>
Subject: [jira] [Commented] (HBASE-6524) Hooks for hbase tracing
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HBASE-6524?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13438391#comment-13438391 ] 

Jonathan Leavitt commented on HBASE-6524:
-----------------------------------------

@Stack
bq. Is htrace up in mvn central?
Yes

{quote} Why do we create a new instance of TInfo here rather than just use the one we have (presuming hasTinfo means the request has a TInfo?)?
+ call = new Call(id, param, this, responder, callSize, new TraceInfo(
+ request.getTinfo().getTraceId(), request.getTinfo().getParentId()));{quote}
I'm trying to keep HTrace agnostic of RPC mechanism, so the TraceInfo in the request is a protobuf TraceInfo (RPCTinfo from Tracing.proto). I just convert it to the TraceInfo defined in HTrace to decrease dependency on protobufs. 
On the other hand, it's not a big deal to store the RPCTinfo in the Call, so if you think the extra overhead from crating the TraceInfo is bad, I'm happy to change it. 

bq. Should say what a spanReceiverHost is. When I load receivers, what kind of resources are we allocating?
Do you mean I should explain SpanReceiverHosts in HMaster? Or describe them more thoroughly in SpanReceiverHost.java? Or somewhere else? 

bq. Is that all it takes to add tracing? That is pretty sweet. I'd say this patch is close to commit. See comments above. You going to do a bit of a blog w/ fancy tracing pictures? Something we could reference from the refguide? Any more pretty pictures for us?
This patch is enough to add tracing for RPCs. There are still places in HBase that need to be instrumented (by that I mean wrapping any thread changes).  It would also be useful to add some more spans for interesting HBaseEvents (maybe start a new span any time we talk to HDFS). 

I will work on a blog post.  I started working on a more thorough README for [htrace|https://github.com/cloudera/htrace]. After I am happy with that I can write up something more specific for the refguide if necessary. 

                
> Hooks for hbase tracing
> -----------------------
>
>                 Key: HBASE-6524
>                 URL: https://issues.apache.org/jira/browse/HBASE-6524
>             Project: HBase
>          Issue Type: Sub-task
>            Reporter: Jonathan Leavitt
>         Attachments: hbase-6524.diff
>
>
> Includes the hooks that use [htrace|http://www.github.com/cloudera/htrace] library to add dapper-like tracing to hbase.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira