[
https://issues.apache.org/jira/browse/CASSANDRA-5077?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Yuki Morishita reopened CASSANDRA-5077:
---------------------------------------
This broke CompactionsPurgeTest because if the last iterated row in compaction
task is empty, SSTableWriter does not get closed.
The code below fixes this(but I feel a bit redundant):
{code}
AbstractCompactedRow row = iter.next();
if (row.isEmpty())
{
controller.invalidateCachedRow(row.key);
row.close();
+ // make sure we close SSTableWriter at the end of iteration
+ if (!iter.hasNext())
+ {
+ SSTableReader toIndex =
writer.closeAndOpenReader(getMaxDataAge(toCompact));
+ cachedKeyMap.put(toIndex, cachedKeys);
+ sstables.add(toIndex);
+ }
continue;
}
{code}
> Simplify CompactionIterable.getReduced
> --------------------------------------
>
> Key: CASSANDRA-5077
> URL: https://issues.apache.org/jira/browse/CASSANDRA-5077
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Jonathan Ellis
> Assignee: Jonathan Ellis
> Fix For: 1.2.1
>
> Attachments: 5077.txt, 5077.txt
>
>
> CompactionIterable turns empty rows into null, and
> purge-compacted-rows-from-cache logic is duplicated (and different!) across
> implementations. Let's simplify to just return empty rows as-is and let
> CompactionTask handle the purge logic.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
