[
https://issues.apache.org/jira/browse/HBASE-14873?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15037212#comment-15037212
]
ramkrishna.s.vasudevan commented on HBASE-14873:
------------------------------------------------
Interesting and useful information here.
bq.DirectBuffer's override method calls Unsafe's copyMemory under the limit
Bits.UNSAFE_COPY_THRESHOLD, which has been bq.introduced for scaling well,
according to google. How copyFromBufferToArray can omit the logic? Another
thing I should bq.mention is that copyFromBufferToArray is forced to use
one-by-one copy for the other VM implementations, which is an almost bq.bug.
Yes true. But our intention was to mainly avoid all those boundary checks that
Oracle's impl has. Some of the earlier tests suggested that the NIO buffer APIs
were not allowing the JVM to do proper inlining because of the boundary checks
in every API it has. the same is true with the onheap BBs also. The aim of
these Byte and BytebufferUtils was to avoid these and directly go with Unsafe.
One thing to see is that if all the are all onheap we just treat them as Arrays
But the problems you seem to say needs to handled in the BBUtils impl? If so,
that seems to be an important thing. But our code handles the endian thing -
then may be that is not enough.
bq.In addition, Oracle seems to plan to remove Unsafe in Java 9.
This as we know they may remove but I think they are planning to provide some
other way of using them? There was a study internally conducted and it was
found that many of the projects (apache projects included) make use of these
DBBs and Unsafe Utils quite extensively.
> Problems around BoundedByteBufferPool providing direct buffers
> --------------------------------------------------------------
>
> Key: HBASE-14873
> URL: https://issues.apache.org/jira/browse/HBASE-14873
> Project: HBase
> Issue Type: Bug
> Reporter: Hiroshi Ikeda
> Assignee: Hiroshi Ikeda
> Attachments: HBASE-14873-V2.patch, HBASE-14873.patch,
> HBASE-14873.patch, HBASE-14873.patch
>
>
> HBASE-13819 made BoundedByteBufferPool provide direct buffers.
> See RpcServer.java:
> {code}
> ...
> class Call implements RpcCallContext {
> protected synchronized void setResponse(...) {
> ...
> this.cellBlock = ipcUtil.buildCellBlock(..., reservoir);
> ...
> bc = new BufferChain(..., this.cellBlock);
> if (connection.useWrap) {
> bc = wrapWithSasl(bc);
> }
> ...
> private BufferChain wrapWithSasl(BufferChain bc) throws IOException {
> ...
> byte[] responseBytes = bc.getBytes();
> ...
> {code}
> {{cellBlock}} is expected to be a direct buffer retrieved from {{reservoir}}
> (but not always), and {{bc}} may be composed of both direct and non-direct
> buffers.
> And then, see BufferChain.java:
> {code}
> byte [] getBytes() {
> ...
> for (ByteBuffer bb: this.buffers) {
> System.arraycopy(bb.array(), ...);
> {code}
> A direct buffer doesn't give its array, and will throw
> UnsupportedOperationException.
> Another problem; {{cellBlock}} is allowed to be a non-direct buffer, and
> after use it will be put to {{reservoir}}, mixing direct and non-direct
> buffers in the pool.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
