Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

HADOOP-18756. S3A prefetch - CachingBlockManager to use AtomicBoolean for closed flag #5718

Merged
merged 4 commits into from
Jun 14, 2023
Merged
Changes from 3 commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Original file line number Diff line number Diff line change
Expand Up @@ -38,6 +38,7 @@
import java.util.Set;
import java.util.concurrent.ConcurrentHashMap;
import java.util.concurrent.TimeUnit;
import java.util.concurrent.atomic.AtomicBoolean;
import java.util.concurrent.locks.ReentrantReadWriteLock;

import org.apache.hadoop.thirdparty.com.google.common.collect.ImmutableSet;
Expand Down Expand Up @@ -68,7 +69,7 @@ public class SingleFilePerBlockCache implements BlockCache {
*/
private int numGets = 0;

private boolean closed;
private final AtomicBoolean closed;

private final PrefetchingStatistics prefetchingStatistics;

Expand Down Expand Up @@ -174,6 +175,7 @@ private boolean takeLock(LockType lockType, long timeout, TimeUnit unit) {
*/
public SingleFilePerBlockCache(PrefetchingStatistics prefetchingStatistics) {
this.prefetchingStatistics = requireNonNull(prefetchingStatistics);
this.closed = new AtomicBoolean(false);
}

/**
Expand Down Expand Up @@ -207,7 +209,7 @@ public int size() {
*/
@Override
public void get(int blockNumber, ByteBuffer buffer) throws IOException {
if (closed) {
if (closed.get()) {
return;
}

Expand Down Expand Up @@ -262,7 +264,7 @@ private Entry getEntry(int blockNumber) {
@Override
public void put(int blockNumber, ByteBuffer buffer, Configuration conf,
LocalDirAllocator localDirAllocator) throws IOException {
if (closed) {
if (closed.get()) {
return;
}

Expand Down Expand Up @@ -333,37 +335,31 @@ protected Path getCacheFilePath(final Configuration conf,

@Override
public void close() throws IOException {
if (closed) {
return;
}

closed = true;
if (closed.compareAndSet(false, true)) {
LOG.debug(getStats());
int numFilesDeleted = 0;

LOG.info(getStats());
int numFilesDeleted = 0;

for (Entry entry : blocks.values()) {
boolean lockAcquired = entry.takeLock(Entry.LockType.WRITE, PREFETCH_WRITE_LOCK_TIMEOUT,
PREFETCH_WRITE_LOCK_TIMEOUT_UNIT);
if (!lockAcquired) {
LOG.error("Cache file {} deletion would not be attempted as write lock could not"
+ " be acquired within {} {}", entry.path, PREFETCH_WRITE_LOCK_TIMEOUT,
for (Entry entry : blocks.values()) {
boolean lockAcquired = entry.takeLock(Entry.LockType.WRITE, PREFETCH_WRITE_LOCK_TIMEOUT,
PREFETCH_WRITE_LOCK_TIMEOUT_UNIT);
continue;
}
try {
Files.deleteIfExists(entry.path);
prefetchingStatistics.blockRemovedFromFileCache();
numFilesDeleted++;
} catch (IOException e) {
LOG.debug("Failed to delete cache file {}", entry.path, e);
} finally {
entry.releaseLock(Entry.LockType.WRITE);
if (!lockAcquired) {
LOG.error("Cache file {} deletion would not be attempted as write lock could not"
+ " be acquired within {} {}", entry.path, PREFETCH_WRITE_LOCK_TIMEOUT,
PREFETCH_WRITE_LOCK_TIMEOUT_UNIT);
continue;
}
try {
Files.deleteIfExists(entry.path);
prefetchingStatistics.blockRemovedFromFileCache();
numFilesDeleted++;
} catch (IOException e) {
LOG.error("Failed to delete cache file {}", entry.path, e);
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ok, so this is moving from debug to error. why so?

if we do want to print this, it's not an error. at worst a warn.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

sure let me make it at least warn, this is likely going to be rare event and highlighting it would be helpful debugging stale file records in future.

Copy link
Contributor Author

@virajjasani virajjasani Jun 12, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

with this new addendum (log level change: ERROR to WARN), i re-ran the test suite again (without scale profile, since the addendum change is trivial)

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

re-tested the latest state of the PR with mvn clean verify -Dparallel-tests -DtestsThreadCount=8 -Dscale -Dprefetch against us-west-2

} finally {
entry.releaseLock(Entry.LockType.WRITE);
}
}
}

if (numFilesDeleted > 0) {
LOG.info("Deleted {} cache files", numFilesDeleted);
LOG.debug("Prefetch cache close: Deleted {} cache files", numFilesDeleted);
}
}

Expand Down