[
https://issues.apache.org/jira/browse/CASSANDRA-7694?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14088030#comment-14088030
]
Jonathan Ellis commented on CASSANDRA-7694:
-------------------------------------------
I'm not sure what's going on here, because the log in CASSANDRA-7706 clearly
shows some interruptions being (correctly) logged at info and debug, but others
at error.
CompactionExecutor attempts to deal with CIE as follows:
{code}
I think we have two problems.
CompactionTask assumes it will be run on a CompactionExecutor, which handles
CIE as follows:
{code}
// modified from DebuggableThreadPoolExecutor so that
CompactionInterruptedExceptions are not logged
@Override
public void afterExecute(Runnable r, Throwable t)
{
super.afterExecute(r, t);
if (t == null)
t = DebuggableThreadPoolExecutor.extractThrowable(r);
if (t != null)
{
if (t instanceof CompactionInterruptedException)
{
logger.info(t.getMessage());
logger.debug("Full interruption stack trace:", t);
}
else
{
DebuggableThreadPoolExecutor.handleOrLog(t);
}
}
}
{code}
If [~mshuler] can reproduce with the split test, what does it bisect to?
> Expected Compaction Interruption is logged as ERROR
> ---------------------------------------------------
>
> Key: CASSANDRA-7694
> URL: https://issues.apache.org/jira/browse/CASSANDRA-7694
> Project: Cassandra
> Issue Type: Bug
> Components: Core
> Environment: OSX and Ubuntu 14.04
> Reporter: Philip Thompson
> Assignee: Marcus Eriksson
> Priority: Minor
> Fix For: 2.1.1
>
> Attachments: compaction_error.log
>
>
> As seen in the attached log, occasionally a major compaction will interrupt
> other running compactions. This is not an error and is expected behavior.
> However this is logged at ERROR.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
