[ 
https://issues.apache.org/jira/browse/HADOOP-3844?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12647615#action_12647615
 ] 

Steve Loughran commented on HADOOP-3844:
----------------------------------------

I have a test that brings up a task tracker without a job tracker. I expected 
it to spin in the started and not live state, but instead we arent spinning on 
connection failure, and this change -mine- is the cause; the connect and socket 
exceptions are getting better text, but they are being downgraded at the same 
time.

Propose: connect and socket exceptions are caught and rethrown with new 
exceptions of the same type that include diagnostics; all other types are 
passed through unchanged. 

Given this change makes cluster startup more brittle, I would propose either 
reverting this change or waiting for the replacement. 

Stack trace

java.io.IOException: Call to localhost/127.0.0.1:8012 failed on local 
exception: java.net.ConnectException: Connection refused
at org.apache.hadoop.ipc.Client.call(Client.java:699)
at org.apache.hadoop.ipc.RPC$Invoker.invoke(RPC.java:216)
at org.apache.hadoop.mapred.$Proxy7.getProtocolVersion(Unknown Source)
at org.apache.hadoop.ipc.RPC.getProxy(RPC.java:319)
at org.apache.hadoop.ipc.RPC.getProxy(RPC.java:306)
at org.apache.hadoop.ipc.RPC.getProxy(RPC.java:343)
at org.apache.hadoop.ipc.RPC.waitForProxy(RPC.java:288)
at org.apache.hadoop.mapred.TaskTracker.initialize(TaskTracker.java:453)
at org.apache.hadoop.mapred.TaskTracker.innerStart(TaskTracker.java:831)
at org.apache.hadoop.util.Service.start(Service.java:186)


> include message of local exception in Client call failures
> ----------------------------------------------------------
>
>                 Key: HADOOP-3844
>                 URL: https://issues.apache.org/jira/browse/HADOOP-3844
>             Project: Hadoop Core
>          Issue Type: Improvement
>          Components: ipc
>    Affects Versions: 0.19.0
>            Reporter: Steve Loughran
>            Assignee: Steve Loughran
>            Priority: Minor
>             Fix For: 0.19.0
>
>         Attachments: hadoop-3844.patch, hadoop-3844.patch
>
>
> When Client fails with a local exception, that exception is retained, but the 
> message is not propagated to the new exception text, which becomes simply  
> "Call failed on local exception"
> The forthcoming patch will change such messages to include that of the nested 
> exception, so producing test reports containing useful data such as
> java.io.IOException: Call failed on local exception: Connection refused

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to