accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Christopher Tubbs (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (ACCUMULO-2950) Fix handling of server-side exceptions
Date Tue, 24 Jun 2014 23:08:25 GMT

    [ https://issues.apache.org/jira/browse/ACCUMULO-2950?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14042819#comment-14042819
] 

Christopher Tubbs commented on ACCUMULO-2950:
---------------------------------------------

If somebody can figure out how to wrap our exceptions with something that actually gets to
the ProcessFunction's relevant code as a TException, I'd love to hear it. I tried a bit, and
couldn't make it happen, but this code is very unfamiliar to me. [~ecn] suggested wrapping
Throwable in TimedProcessor.process since it was removed from ProcessFunction in THRIFT-1805,
but that gets eaten in FrameBuffer.invoke, so, I'm pretty sure they've neutered that entire
block of code. I'm not sure how it is expected to get server-side exceptions to not be completely
eaten and seen as network problems (TTransportException).

> Fix handling of server-side exceptions
> --------------------------------------
>
>                 Key: ACCUMULO-2950
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-2950
>             Project: Accumulo
>          Issue Type: Sub-task
>            Reporter: Christopher Tubbs
>            Assignee: Christopher Tubbs
>             Fix For: 1.6.1, 1.7.0
>
>
> THRIFT-1805 caused a second regression of the issue in 0.9.1 that was fixed the first
time by THRIFT-1658 in 0.9.0.
> We need to wrap RuntimeExceptions that can occur on the server side with something that
doesn't kill the connection and result in the client seeing a TTransportException.



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Mime
View raw message