accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mike Drob (JIRA)" <>
Subject [jira] [Commented] (ACCUMULO-2660) InvocationTargetException during ContinuousIngest
Date Fri, 11 Apr 2014 17:04:16 GMT


Mike Drob commented on ACCUMULO-2660:

It still propagates an IOException when that is the underlying cause. In the wrapped RuntimeException
and Error case, we were actively doing the wrong thing, but our current code handles it anyway
(because we continue to do the wrong thing in other places).

{{defineTablet}} is only called from inside of {{write}} of an anonymous {{Writer}} created
in {{TabletServerLogger.defineTablet}}. This {{write}} method is called from {{TabletServerLogger.write}}
(line 275) inside of a {{try}} block that catches {{Exception}} and continues to retry.

In the {{Error}} case, we probably want to stop retrying anyway and let that bubble up, since
it indicates something seriously wrong. We wouldn't be catching it anyway if we were using
a compiled call instead of our reflective one, so there's not a good reason to wrap that.


> InvocationTargetException during ContinuousIngest
> -------------------------------------------------
>                 Key: ACCUMULO-2660
>                 URL:
>             Project: Accumulo
>          Issue Type: Bug
>          Components: tserver
>    Affects Versions: 1.6.0
>         Environment: 5 tserver cluster, cdh4.5
>            Reporter: Mike Drob
>            Assignee: Mike Drob
>              Labels: 16_qa_bug
>             Fix For: 1.6.0
>         Attachments: ACCUMULO-2660.patch.txt
> Two tservers logged the following message during continuous ingest, at different times
(about 4 hours apart).
> {noformat}
> 2014-04-10 22:16:34,220 [log.DfsLogger] ERROR: java.lang.reflect.InvocationTargetException
> {noformat}
> The tservers continued to run, but there was no full trace even at the debug level.

This message was sent by Atlassian JIRA

View raw message