openinx commented on a change in pull request #257: HBASE-22463 Some paths in
HFileScannerImpl did not consider block#release which will exhaust the
ByteBuffAllocator
URL: https://github.com/apache/hbase/pull/257#discussion_r288852974
##########
File path:
hbase-server/src/main/java/org/apache/hadoop/hbase/io/hfile/HFileReaderImpl.java
##########
@@ -865,29 +861,39 @@ public boolean seekBefore(Cell key) throws IOException {
// The key we are interested in
if (previousBlockOffset == -1) {
// we have a 'problem', the key we want is the first of the file.
+ releaseIfNotCurBlock(seekToBlock);
return false;
}
// The first key in the current block 'seekToBlock' is greater than
the given
// seekBefore key. We will go ahead by reading the next block that
satisfies the
// given key. Return the current block before reading the next one.
- seekToBlock.release();
Review comment:
Yeah, similar to the issue:
https://issues.apache.org/jira/browse/HBASE-22480. Discussed with binlijin
before, his patch can apply to branch-2.x and need not to apply to master
branch, once the HBASE-21879 get merged into master, then this bug will be
fixed in this patch.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
