drill-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Daniel Barclay (Drill) (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (DRILL-3322) Something broken in or around RPC timeout setup?
Date Wed, 19 Aug 2015 22:46:45 GMT

    [ https://issues.apache.org/jira/browse/DRILL-3322?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14703867#comment-14703867
] 

Daniel Barclay (Drill) commented on DRILL-3322:
-----------------------------------------------

The point was that it seemed that there might be a latent race-condition bug, one that that
caused timeouts to not send notification to listeners (whatever would have let execute(...)
return or throw a timeout or hey-my-connection's-closed exception).

(Things may have changed since then.)




> Something broken in or around RPC timeout setup?
> ------------------------------------------------
>
>                 Key: DRILL-3322
>                 URL: https://issues.apache.org/jira/browse/DRILL-3322
>             Project: Apache Drill
>          Issue Type: Bug
>          Components: Execution - RPC
>    Affects Versions: 1.0.0
>            Reporter: Daniel Barclay (Drill)
>            Assignee: Jacques Nadeau
>             Fix For: 1.2.0
>
>
> Something(s) in the overall setup for RPC timeouts (heartbeats/watchdog) seems broken.
> Running in the debugger (setting breakpoints, lettings things time out, etc.) has led
to a state where the JDBC layer's RPC client code never gets any callbacks (not even an exception
callback), and so the JDBC execute..(...) call never returns or even throws.



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

Mime
View raw message