[
https://issues.apache.org/jira/browse/HADOOP-18190?focusedWorklogId=792296&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-792296
]
ASF GitHub Bot logged work on HADOOP-18190:
-------------------------------------------
Author: ASF GitHub Bot
Created on: 18/Jul/22 18:11
Start Date: 18/Jul/22 18:11
Worklog Time Spent: 10m
Work Description: steveloughran commented on code in PR #4458:
URL: https://github.com/apache/hadoop/pull/4458#discussion_r923663742
##########
hadoop-common-project/hadoop-common/src/main/java/org/apache/hadoop/fs/statistics/StreamStatisticNames.java:
##########
@@ -387,6 +387,46 @@ public final class StreamStatisticNames {
public static final String BLOCKS_RELEASED
= "blocks_released";
+ /**
+ * Total number of prefetching operations executed.
+ */
+ public static final String STREAM_READ_PREFETCH_OPERATIONS
+ = "stream_read_prefetch_operations";
+
+ /**
+ * Total number of block in disk cache.
+ */
+ public static final String STREAM_READ_BLOCKS_IN_FILE_CACHE
+ = "stream_read_blocks_in_cache";
+
+ /**
+ * Total number of active prefetch operations.
+ */
+ public static final String STREAM_READ_ACTIVE_PREFETCH_OPERATIONS
+ = "stream_read_active_prefetch_operations";
+
+ /**
+ * Total bytes of memory in use by this input stream.
+ */
+ public static final String STREAM_READ_ACTIVE_MEMORY_IN_USE
+ = "stream_read_active_memory_in_use";
+
+ /**
+ * count/duration of reading a remote block.
+ * IO.
Review Comment:
why the IO here and below?
##########
hadoop-tools/hadoop-aws/src/main/java/org/apache/hadoop/fs/common/CachingBlockManager.java:
##########
@@ -330,6 +341,10 @@ private void readBlock(BufferData data, boolean
isPrefetch, BufferData.State...
this.read(buffer, offset, size);
buffer.flip();
data.setReady(expectedState);
+
+ if(isPrefetch) {
Review Comment:
nit, spacing
##########
hadoop-tools/hadoop-aws/src/main/java/org/apache/hadoop/fs/common/BufferPool.java:
##########
@@ -236,11 +244,15 @@ public synchronized void close() {
}
}
+ int currentPoolSize = this.pool.numCreated();
Review Comment:
how about you cut the `this.` here and from the existing calls in the same
method. not our project's style
##########
hadoop-tools/hadoop-aws/src/main/java/org/apache/hadoop/fs/common/CachingBlockManager.java:
##########
@@ -350,16 +365,18 @@ private void readBlock(BufferData data, boolean
isPrefetch, BufferData.State...
private static class PrefetchTask implements Supplier<Void> {
private final BufferData data;
private final CachingBlockManager blockManager;
+ private final Instant taskQueuedStartTime;
- PrefetchTask(BufferData data, CachingBlockManager blockManager) {
+ PrefetchTask(BufferData data, CachingBlockManager blockManager, Instant
taskQueuedStartTime) {
this.data = data;
this.blockManager = blockManager;
+ this.taskQueuedStartTime = taskQueuedStartTime;
}
@Override
public Void get() {
try {
- this.blockManager.prefetch(data);
+ this.blockManager.prefetch(data, taskQueuedStartTime);
} catch (Exception e) {
LOG.error("error during prefetch", e);
Review Comment:
i worry about this. seen problems with transient network failures and abfs
where the logs are flooded with many parallel stack traces before it recovers.
##########
hadoop-tools/hadoop-aws/src/main/java/org/apache/hadoop/fs/common/CachingBlockManager.java:
##########
@@ -330,6 +341,10 @@ private void readBlock(BufferData data, boolean
isPrefetch, BufferData.State...
this.read(buffer, offset, size);
buffer.flip();
data.setReady(expectedState);
+
+ if(isPrefetch) {
+ this.prefetchingStatistics.prefetchOperationCompleted();
Review Comment:
what happens if the read failed? is the failure count and duration before
failure recorded separately from the successful calls?
##########
hadoop-tools/hadoop-aws/src/main/java/org/apache/hadoop/fs/common/BufferPool.java:
##########
@@ -56,26 +58,32 @@ public class BufferPool implements Closeable {
// Allows associating metadata to each buffer in the pool.
private Map<BufferData, ByteBuffer> allocated;
+ private PrefetchingStatistics prefetchingStatistics;
+
/**
* Initializes a new instance of the {@code BufferPool} class.
*
* @param size number of buffer in this pool.
* @param bufferSize size in bytes of each buffer.
+ * @param prefetchingStatistics statistics for this stream.
*
* @throws IllegalArgumentException if size is zero or negative.
* @throws IllegalArgumentException if bufferSize is zero or negative.
*/
- public BufferPool(int size, int bufferSize) {
+ public BufferPool(int size, int bufferSize, PrefetchingStatistics
prefetchingStatistics) {
Review Comment:
we are going to have to rename this class;
org.apache.hadoop.io.ByteBufferPool has long existed and it will only cause
confusion. not in this PR, but soon
Issue Time Tracking
-------------------
Worklog Id: (was: 792296)
Time Spent: 1.5h (was: 1h 20m)
> s3a prefetching streams to collect iostats on prefetching operations
> --------------------------------------------------------------------
>
> Key: HADOOP-18190
> URL: https://issues.apache.org/jira/browse/HADOOP-18190
> Project: Hadoop Common
> Issue Type: Sub-task
> Components: fs/s3
> Affects Versions: 3.4.0
> Reporter: Steve Loughran
> Assignee: Ahmar Suhail
> Priority: Major
> Labels: pull-request-available
> Time Spent: 1.5h
> Remaining Estimate: 0h
>
> There is a lot more happening in reads, so there's a lot more data to collect
> and publish in IO stats for us to view in a summary at the end of processes
> as well as get from the stream while it is active.
> Some useful ones would seem to be:
> counters
> * is in memory. using 0 or 1 here lets aggregation reports count total #of
> memory cached files.
> * prefetching operations executed
> * errors during prefetching
> gauges
> * number of blocks in cache
> * total size of blocks
> * active prefetches
> + active memory used
> duration tracking count/min/max/ave
> * time to fetch a block
> * time queued before the actual fetch begins
> * time a reader is blocked waiting for a block fetch to complete
> and some info on cache use itself
> * number of blocks discarded unread
> * number of prefetched blocks later used
> * number of backward seeks to a prefetched block
> * number of forward seeks to a prefetched block
> the key ones I care about are
> # memory consumption
> # can we determine if cache is working (reads with cache hit) and when it is
> not (misses, wasted prefetches)
> # time blocked on executors
> The stats need to be accessible on a stream even when closed, and aggregated
> into the FS. once we get per-thread stats contexts we can publish there too
> and collect in worker threads for reporting in task commits
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
