[
https://issues.apache.org/jira/browse/HADOOP-19394?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18010911#comment-18010911
]
ASF GitHub Bot commented on HADOOP-19394:
-----------------------------------------
ahmarsuhail commented on code in PR #7720:
URL: https://github.com/apache/hadoop/pull/7720#discussion_r2242680859
##########
hadoop-tools/hadoop-aws/src/main/java/org/apache/hadoop/fs/s3a/impl/streams/AnalyticsStream.java:
##########
@@ -128,6 +141,48 @@ public int read(byte[] buf, int off, int len) throws
IOException {
return bytesRead;
}
+ /**
+ * {@inheritDoc}
+ * Pass to {@link #readVectored(List, IntFunction, Consumer)}
+ * with the {@link VectoredReadUtils#LOG_BYTE_BUFFER_RELEASED} releaser.
+ * @param ranges the byte ranges to read.
+ * @param allocate the function to allocate ByteBuffer.
+ * @throws IOException IOE if any.
+ */
+ @Override
+ public void readVectored(List<? extends FileRange> ranges,
+ IntFunction<ByteBuffer> allocate)
throws IOException {
+ readVectored(ranges, allocate, LOG_BYTE_BUFFER_RELEASED);
+ }
+
+ /**
+ * {@inheritDoc}
+ * Pass to {@link #readVectored(List, IntFunction, Consumer)}
+ * with the {@link VectoredReadUtils#LOG_BYTE_BUFFER_RELEASED} releaser.
+ * @param ranges the byte ranges to read.
+ * @param allocate the function to allocate ByteBuffer.
+ * @throws IOException IOE if any.
+ */
+ @Override
+ public void readVectored(final List<? extends FileRange> ranges,
+ final IntFunction<ByteBuffer> allocate,
+ final Consumer<ByteBuffer> release) throws
IOException {
+ LOG.debug("AAL: Starting vectored read on path {} for ranges {} ",
getPathStr(), ranges);
+ throwIfClosed();
+
+ List<ObjectRange> objectRanges = new ArrayList<>();
+
+ for (FileRange range : ranges) {
+ CompletableFuture<ByteBuffer> result = new CompletableFuture<>();
+ ObjectRange objectRange = new ObjectRange(result, range.getOffset(),
range.getLength());
+ objectRanges.add(objectRange);
+ range.setData(result);
+ }
+
+ // AAL does not do any range coalescing, so input and combined ranges are
the same.
+ this.getS3AStreamStatistics().readVectoredOperationStarted(ranges.size(),
ranges.size());
+ inputStream.readVectored(objectRanges, allocate, release);
Review Comment:
ah no :( release isn't called right now, but will make that change and get
it released in the next version
> S3A Analytics Accelerator: vector IO support
> --------------------------------------------
>
> Key: HADOOP-19394
> URL: https://issues.apache.org/jira/browse/HADOOP-19394
> Project: Hadoop Common
> Issue Type: Sub-task
> Components: fs/s3
> Affects Versions: 3.4.1
> Reporter: Steve Loughran
> Priority: Major
> Labels: pull-request-available
>
> Add vector IO support for analytics accelerator stream
> Three stages
> # pull up s3a input stream to work with all ObjectInputStreams; do its own
> fetching independent of the analytics stream
> # provide info to stream of fetches having taken place (remove from cache,
> cancel prefetch)
> full integration
> * return a range from cache if present
> * append to the block retrieval callback if a prefetch is in progress
> * only do merge + new request if the range cannot be satisifed entirely from
> cached data
> Out of scope: handling case where part of a range is in cache/retrieval. Too
> complicated and so prone to problems.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: common-issues-h...@hadoop.apache.org
