[ https://issues.apache.org/jira/browse/STORM-87?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14049143#comment-14049143
]
ASF GitHub Bot commented on STORM-87:
-------------------------------------
Github user revans2 commented on a diff in the pull request:
https://github.com/apache/incubator-storm/pull/46#discussion_r14420063
--- Diff: storm-core/src/jvm/backtype/storm/task/ShellBolt.java ---
@@ -241,5 +241,8 @@ private void handleEmit(Map action) throws InterruptedException {
private void die(Throwable exception) {
_exception = exception;
+ LOG.info("Halting process: ShellBolt died.", exception);
+ _collector.reportError(exception);
+ Runtime.getRuntime().halt(11);
--- End diff --
Now that we have ways of cleanly handling System.exit, can we switch over to that instead
of Runtime.halt?
> ShellBolt shoud fail fast on subprocess error
> ---------------------------------------------
>
> Key: STORM-87
> URL: https://issues.apache.org/jira/browse/STORM-87
> Project: Apache Storm (Incubating)
> Issue Type: Bug
> Reporter: James Xu
> Priority: Minor
>
> https://github.com/nathanmarz/storm/issues/542
> Now ShellBolt.die() just save the catched exception and throw it in next execute() call.
If a tuple cause subprocess fail and next tuple received after some time, the ShellBolt will
fail until next tuple. It maybe more useful to exit immediately in die().
--
This message was sent by Atlassian JIRA
(v6.2#6252)
|