[
https://issues.apache.org/jira/browse/HBASE-7947?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13588974#comment-13588974
]
Jesse Yates commented on HBASE-7947:
------------------------------------
Nope, looks like its actually a bug. The relevant code is from
CompactSplitThread, lines 206-233:
{code}
@Override
public synchronized List<CompactionRequest> requestCompaction(final HRegion
r, final String why,
int p, List<CompactionRequest> requests) throws IOException {
// not a special compaction request, so make our own list
List<CompactionRequest> ret;
if (requests == null) {
ret = new ArrayList<CompactionRequest>(r.getStores().size());
for (Store s : r.getStores().values()) {
ret.add(requestCompaction(r, s, why, p, null));
}
} else {
ret = new ArrayList<CompactionRequest>(requests.size());
for (CompactionRequest request : requests) {
requests.add(requestCompaction(r, request.getStore(), why, p, request));
}
}
return ret;
}
{code}
Patch coming momentarily
> Custom compaction requests should have test for adding to wrong list
> --------------------------------------------------------------------
>
> Key: HBASE-7947
> URL: https://issues.apache.org/jira/browse/HBASE-7947
> Project: HBase
> Issue Type: Bug
> Reporter: Jesse Yates
> Assignee: Jesse Yates
> Fix For: 0.95.0, 0.96.0
>
>
> This came from backporting HBASE-7725 to 0.94. Essentially, the created
> compaction requests were added to the wrong list, resulting in concurrent
> modifications exceptions and incorrect return values when used in practice.
> The code is fine in 0.95/trunk, but should include the same added test to
> catch a possible regression.
--
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
