You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-issues@hadoop.apache.org by GitBox <gi...@apache.org> on 2020/03/17 04:37:11 UTC
[GitHub] [hadoop] snvijaya opened a new pull request #1898: Report
read-ahead error back
snvijaya opened a new pull request #1898: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898
Currently errors in read-ahead are silently ignored, thus failing to highlight any issues and causing slowness to the overall read request.
Any new read request in-turn triggers n num of read-aheads and all of them will silently fail.
This PR will report back error from the read-ahead issued by the active read call. Also, cause subsequent reads to only retry the respective read position based on the failure seen for the previous read-ahead failure on same position.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] snvijaya commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
snvijaya commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r394940399
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/AbfsInputStream.java
##########
@@ -234,6 +242,7 @@ int readRemote(long position, byte[] b, int offset, int length) throws IOExcepti
final AbfsRestOperation op;
AbfsPerfTracker tracker = client.getAbfsPerfTracker();
try (AbfsPerfInfo perfInfo = new AbfsPerfInfo(tracker, "readRemote", "read")) {
+ LOG.trace(String.format("Trigger client.read for path=%s position=%s offset=%s length=%s", path, position, offset, length));
Review comment:
Done
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] steveloughran commented on issue #1898: HADOOP-16852:
Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
steveloughran commented on issue #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#issuecomment-609770402
@DadanielZ is happy with the core patch, so I am too. just the checkstyle to fix
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] snvijaya commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
snvijaya commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r400112122
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBufferManager.java
##########
@@ -299,11 +327,32 @@ private void clearFromReadAheadQueue(final AbfsInputStream stream, final long re
}
private int getBlockFromCompletedQueue(final AbfsInputStream stream, final long position, final int length,
- final byte[] buffer) {
- ReadBuffer buf = getFromList(completedReadList, stream, position);
- if (buf == null || position >= buf.getOffset() + buf.getLength()) {
+ final byte[] buffer) throws IOException {
+ ReadBuffer buf = getBufferFromCompletedQueue(stream, position);
+
+ if (buf == null) {
return 0;
}
+
+ if (buf.getStatus() == ReadBufferStatus.READ_FAILED) {
+ // Eviction of a read buffer is triggered only when a queue request comes in
+ // and each eviction attempt tries to find one eligible buffer.
+ // Hence there are chances that an old read-ahead buffer with exception is still
+ // available. To prevent new read requests to fail due to such old buffers,
+ // return exception only from buffers that failed within last THRESHOLD_AGE_MILLISECONDS
+ if ((currentTimeMillis() - (buf.getTimeStamp()) < THRESHOLD_AGE_MILLISECONDS)) {
Review comment:
Aim here is to enforce the read-ahead failure for the threshold time duration (which currently is 30 sec), i.e. any read request for that offset that can be served by the ReadBuffer needs to fail.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] snvijaya commented on issue #1898: HADOOP-16852: Report
read-ahead error back
Posted by GitBox <gi...@apache.org>.
snvijaya commented on issue #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#issuecomment-607010089
> needs a hadoop JIRA and a link back. PRs without a matching JIRA do not exist and SHALL not be committed
Have made the necessary updates.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] snvijaya commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
snvijaya commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r394940802
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBufferManager.java
##########
@@ -90,8 +94,10 @@ private ReadBufferManager() {
* @param stream The {@link AbfsInputStream} for which to do the read-ahead
* @param requestedOffset The offset in the file which shoukd be read
* @param requestedLength The length to read
+ * @param queueReadAheadRequestId unique queue request ID
*/
- void queueReadAhead(final AbfsInputStream stream, final long requestedOffset, final int requestedLength) {
+ void queueReadAhead(final AbfsInputStream stream, final long requestedOffset, final int requestedLength
Review comment:
Undid any change to method signature as part of the review comments to remove queueRequestId.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] goiri commented on a change in pull request #1898: Report
read-ahead error back
Posted by GitBox <gi...@apache.org>.
goiri commented on a change in pull request #1898: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r393790939
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/AbfsInputStream.java
##########
@@ -234,6 +242,7 @@ int readRemote(long position, byte[] b, int offset, int length) throws IOExcepti
final AbfsRestOperation op;
AbfsPerfTracker tracker = client.getAbfsPerfTracker();
try (AbfsPerfInfo perfInfo = new AbfsPerfInfo(tracker, "readRemote", "read")) {
+ LOG.trace(String.format("Trigger client.read for path=%s position=%s offset=%s length=%s", path, position, offset, length));
Review comment:
Can you use logger format?
LOG.trace("Trigger client.read for path={} position={} offset={} length={}", path, position, offset, length);
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] steveloughran commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
steveloughran commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r405032853
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBufferManager.java
##########
@@ -289,6 +298,27 @@ private ReadBuffer getFromList(final Collection<ReadBuffer> list, final AbfsInpu
return null;
}
+ /**
+ * Returns buffers that failed or passed from completed queue
+ * @param stream
+ * @param requestedOffset
+ * @return
+ */
+ private ReadBuffer getBufferFromCompletedQueue(final AbfsInputStream stream, final long requestedOffset) {
+ for (ReadBuffer buffer : completedReadList) {
+ // Buffer is returned if the requestedOffset is at or above buffer's
+ // offset but less than buffer's length or the actual requestedLength
+ if ((buffer.getStream() == stream)
+ && (requestedOffset >= buffer.getOffset())
+ && ((requestedOffset < buffer.getOffset() + buffer.getLength())
Review comment:
Should always be for requested length? That is: we don't care about the total length of the buffer, only the amount which was requested from the far end -which may be less?
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] snvijaya commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
snvijaya commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r400114094
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBuffer.java
##########
@@ -88,12 +92,23 @@ public void setBufferindex(int bufferindex) {
this.bufferindex = bufferindex;
}
+ public IOException getErrException() {
+ return errException;
+ }
+
+ public void setErrException(final java.io.IOException errException) {
Review comment:
Fixed
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] steveloughran commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
steveloughran commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r405031623
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBufferManager.java
##########
@@ -289,6 +298,27 @@ private ReadBuffer getFromList(final Collection<ReadBuffer> list, final AbfsInpu
return null;
}
+ /**
+ * Returns buffers that failed or passed from completed queue
Review comment:
add trailing .
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] goiri commented on a change in pull request #1898: Report
read-ahead error back
Posted by GitBox <gi...@apache.org>.
goiri commented on a change in pull request #1898: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r393791537
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBufferManager.java
##########
@@ -101,6 +107,7 @@ void queueReadAhead(final AbfsInputStream stream, final long requestedOffset, fi
if (isAlreadyQueued(stream, requestedOffset)) {
return; // already queued, do not queue again
}
+
Review comment:
Avoid.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] snvijaya commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
snvijaya commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r394942002
##########
File path: hadoop-tools/hadoop-azure/src/test/java/org/apache/hadoop/fs/azurebfs/services/TestAbfsInputStream.java
##########
@@ -0,0 +1,438 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.hadoop.fs.azurebfs.services;
+
+import java.io.IOException;
+import java.util.UUID;
+
+import org.junit.Assert;
+import org.junit.Test;
+
+import org.apache.hadoop.fs.azurebfs.AbstractAbfsIntegrationTest;
+import org.apache.hadoop.fs.azurebfs.contracts.exceptions.AzureBlobFileSystemException;
+import org.apache.hadoop.fs.azurebfs.contracts.exceptions.TimeoutException;
+
+import static java.util.UUID.randomUUID;
+import static org.mockito.ArgumentMatchers.any;
+import static org.mockito.Mockito.doReturn;
+import static org.mockito.Mockito.doThrow;
+import static org.mockito.Mockito.mock;
+import static org.mockito.Mockito.times;
+import static org.mockito.Mockito.verify;
+import static org.mockito.Mockito.when;
+import static org.apache.hadoop.test.LambdaTestUtils.intercept;
+import static org.apache.hadoop.fs.azurebfs.constants.AbfsHttpConstants.FORWARD_SLASH;
+
+/**
+ * Unit test AbfsInputStream.
+ */
+public class TestAbfsInputStream extends
Review comment:
If you meant capturing the exception messages from failed read ahead threads, that will not be possible as we can not predict which stub hit which of the parallely running read ahead threads. Hence asserting on the exception and no message.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] steveloughran commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
steveloughran commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r405026217
##########
File path: hadoop-tools/hadoop-azure/src/test/java/org/apache/hadoop/fs/azurebfs/services/TestAbfsInputStream.java
##########
@@ -0,0 +1,438 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.hadoop.fs.azurebfs.services;
+
+import java.io.IOException;
+import java.util.UUID;
+
+import org.junit.Assert;
+import org.junit.Test;
+
+import org.apache.hadoop.fs.azurebfs.AbstractAbfsIntegrationTest;
+import org.apache.hadoop.fs.azurebfs.contracts.exceptions.AzureBlobFileSystemException;
+import org.apache.hadoop.fs.azurebfs.contracts.exceptions.TimeoutException;
+
+import static java.util.UUID.randomUUID;
+import static org.mockito.ArgumentMatchers.any;
+import static org.mockito.Mockito.doReturn;
+import static org.mockito.Mockito.doThrow;
+import static org.mockito.Mockito.mock;
+import static org.mockito.Mockito.times;
+import static org.mockito.Mockito.verify;
+import static org.mockito.Mockito.when;
+import static org.apache.hadoop.test.LambdaTestUtils.intercept;
+import static org.apache.hadoop.fs.azurebfs.constants.AbfsHttpConstants.FORWARD_SLASH;
+
+/**
+ * Unit test AbfsInputStream.
+ */
+public class TestAbfsInputStream extends
Review comment:
that's always been trouble in the past; major source of maintenance pain in the wasb tests. Better use the assertion
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] snvijaya commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
snvijaya commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r396233215
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBufferManager.java
##########
@@ -299,11 +327,32 @@ private void clearFromReadAheadQueue(final AbfsInputStream stream, final long re
}
private int getBlockFromCompletedQueue(final AbfsInputStream stream, final long position, final int length,
- final byte[] buffer) {
- ReadBuffer buf = getFromList(completedReadList, stream, position);
- if (buf == null || position >= buf.getOffset() + buf.getLength()) {
+ final byte[] buffer) throws IOException {
+ ReadBuffer buf = getBufferFromCompletedQueue(stream, position);
+
+ if (buf == null) {
return 0;
}
+
+ if (buf.getStatus() == ReadBufferStatus.READ_FAILED) {
+ // Eviction of a read buffer is triggered only when a queue request comes in
+ // and each eviction attempt tries to find one eligible buffer.
+ // Hence there are chances that an old read-ahead buffer with exception is still
+ // available. To prevent new read requests to fail due to such old buffers,
+ // return exception only from buffers that failed within last THRESHOLD_AGE_MILLISECONDS
+ if ((currentTimeMillis() - (buf.getTimeStamp()) < THRESHOLD_AGE_MILLISECONDS)) {
Review comment:
If the buffer was updated with an error in last "Threshold_age_milliseconds", then we return the error to client.
Can you please help me with the cause of confusion.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] DadanielZ commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
DadanielZ commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r399723610
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBufferManager.java
##########
@@ -141,7 +144,8 @@ void queueReadAhead(final AbfsInputStream stream, final long requestedOffset, fi
* @param buffer the buffer to read data into. Note that the buffer will be written into from offset 0.
* @return the number of bytes read
*/
- int getBlock(final AbfsInputStream stream, final long position, final int length, final byte[] buffer) {
+ int getBlock(final AbfsInputStream stream, final long position, final int length, final byte[] buffer)
+ throws java.io.IOException {
Review comment:
why use `java.io.IOException `but not `IOException`?
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] goiri commented on a change in pull request #1898: Report
read-ahead error back
Posted by GitBox <gi...@apache.org>.
goiri commented on a change in pull request #1898: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r393793595
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBufferManager.java
##########
@@ -141,7 +149,8 @@ void queueReadAhead(final AbfsInputStream stream, final long requestedOffset, fi
* @param buffer the buffer to read data into. Note that the buffer will be written into from offset 0.
* @return the number of bytes read
Review comment:
new param
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] steveloughran commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
steveloughran commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r405030875
##########
File path: hadoop-tools/hadoop-azure/src/test/java/org/apache/hadoop/fs/azurebfs/services/TestAbfsInputStream.java
##########
@@ -0,0 +1,433 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.hadoop.fs.azurebfs.services;
+
+import java.io.IOException;
+
+import org.junit.Assert;
+import org.junit.Test;
+
+import org.apache.hadoop.fs.azurebfs.AbstractAbfsIntegrationTest;
+import org.apache.hadoop.fs.azurebfs.contracts.exceptions.AzureBlobFileSystemException;
+import org.apache.hadoop.fs.azurebfs.contracts.exceptions.TimeoutException;
+
+import static org.mockito.ArgumentMatchers.any;
+import static org.mockito.Mockito.doReturn;
+import static org.mockito.Mockito.doThrow;
+import static org.mockito.Mockito.mock;
+import static org.mockito.Mockito.times;
+import static org.mockito.Mockito.verify;
+import static org.mockito.Mockito.when;
+import static org.apache.hadoop.test.LambdaTestUtils.intercept;
+import static org.apache.hadoop.fs.azurebfs.constants.AbfsHttpConstants.FORWARD_SLASH;
+
+/**
+ * Unit test AbfsInputStream.
+ */
+public class TestAbfsInputStream extends
+ AbstractAbfsIntegrationTest {
+
+ private static final int KILOBYTE = 1024;
+
+ private AbfsRestOperation getMockRestOp() {
+ AbfsRestOperation op = mock(AbfsRestOperation.class);
+ AbfsHttpOperation httpOp = mock(AbfsHttpOperation.class);
+ when(httpOp.getBytesReceived()).thenReturn(1024L);
+ when(op.getResult()).thenReturn(httpOp);
+ return op;
+ }
+
+ private AbfsClient getMockAbfsClient() {
+ // Mock failure for client.read()
+ AbfsClient client = mock(AbfsClient.class);
+ AbfsPerfTracker tracker = new AbfsPerfTracker(
+ "test",
+ this.getAccountName(),
+ this.getConfiguration());
+ when(client.getAbfsPerfTracker()).thenReturn(tracker);
+
+ return client;
+ }
+
+ private AbfsInputStream getAbfsInputStream(AbfsClient mockAbfsClient, String fileName) {
+ // Create AbfsInputStream with the client instance
+ AbfsInputStream inputStream = new AbfsInputStream(
+ mockAbfsClient,
+ null,
+ FORWARD_SLASH + fileName,
+ 3 * KILOBYTE,
+ 1 * KILOBYTE, // Setting read ahead buffer size of 1 KB
+ this.getConfiguration().getReadAheadQueueDepth(),
+ this.getConfiguration().getTolerateOobAppends(),
+ "eTag");
+
+ return inputStream;
+ }
+
+ private void queueReadAheads(AbfsInputStream inputStream) {
+ // Mimic AbfsInputStream readAhead queue requests
+ ReadBufferManager.getBufferManager()
+ .queueReadAhead(inputStream, 0, 1 * KILOBYTE);
+ ReadBufferManager.getBufferManager()
+ .queueReadAhead(inputStream, 1 * KILOBYTE, 1 * KILOBYTE);
+ ReadBufferManager.getBufferManager()
+ .queueReadAhead(inputStream, 2 * KILOBYTE, 1 * KILOBYTE);
+ }
+
+ private void verifyReadCallCount(AbfsClient client, int count) throws
+ AzureBlobFileSystemException, InterruptedException {
+ // ReadAhead threads are triggered asynchronously.
+ // Wait a second before verifying the number of total calls.
+ Thread.sleep(1000);
+ verify(client, times(count)).read(any(String.class), any(Long.class),
+ any(byte[].class), any(Integer.class), any(Integer.class),
+ any(String.class));
+ }
+
+ private void checkEvictedStatus(AbfsInputStream inputStream, int position, boolean expectedToThrowException)
+ throws Exception {
+ // Sleep for the eviction threshold time
+ Thread.sleep(ReadBufferManager.getBufferManager().getThresholdAgeMilliseconds() + 1000);
+
+ // Eviction is done only when AbfsInputStream tries to queue new items.
+ // 1 tryEvict will remove 1 eligible item. To ensure that the current test buffer
+ // will get evicted (considering there could be other tests running in parallel),
+ // call tryEvict for the number of items that are there in completedReadList.
+ int numOfCompletedReadListItems = ReadBufferManager.getBufferManager().getCompletedReadListSize();
+ while (numOfCompletedReadListItems > 0) {
+ ReadBufferManager.getBufferManager().callTryEvict();
+ numOfCompletedReadListItems--;
+ }
+
+ if (expectedToThrowException) {
+ intercept(IOException.class,
+ () -> inputStream.read(position, new byte[1 * KILOBYTE], 0, 1 * KILOBYTE));
+ } else {
+ inputStream.read(position, new byte[1 * KILOBYTE], 0, 1 * KILOBYTE);
+ }
+ }
+
+ public TestAbfsInputStream() throws Exception {
+ super();
+ }
+
+ /**
+ * This test expects AbfsInputStream to throw the exception that readAhead
+ * thread received on read. The readAhead thread must be initiated from the
+ * active read request itself.
+ * Also checks that the ReadBuffers are evicted as per the ReadBufferManager
+ * threshold criteria.
+ * @throws Exception
+ */
+ @Test
+ public void testFailedReadAhead() throws Exception {
+ AbfsClient client = getMockAbfsClient();
+ AbfsRestOperation successOp = getMockRestOp();
+
+ // Stub :
+ // Read request leads to 3 readahead calls: Fail all 3 readahead-client.read()
+ // Actual read request fails with the failure in readahead thread
+ doThrow(new TimeoutException("Internal Server error for RAH-Thread-X"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Thread-Y"))
+ .doThrow(new TimeoutException("Internal Server error RAH-Thread-Z"))
+ .doReturn(successOp) // Any extra calls to read, pass it.
+ .when(client)
+ .read(any(String.class), any(Long.class), any(byte[].class),
+ any(Integer.class), any(Integer.class), any(String.class));
+
+ AbfsInputStream inputStream = getAbfsInputStream(client, "testFailedReadAhead.txt");
+
+ // Scenario: ReadAhead triggered from current active read call failed
+ // Before the change to return exception from readahead buffer,
+ // AbfsInputStream would have triggered an extra readremote on noticing
+ // data absent in readahead buffers
+ // In this test, a read should trigger 3 client.read() calls as file is 3 KB
+ // and readahead buffer size set in AbfsInputStream is 1 KB
+ // There should only be a total of 3 client.read() in this test.
+ intercept(IOException.class,
+ () -> inputStream.read(new byte[1 * KILOBYTE]));
+
+ // Only the 3 readAhead threads should have triggered client.read
+ verifyReadCallCount(client, 3);
+
+ // Stub returns success for the 4th read request, if ReadBuffers still
+ // persisted, ReadAheadManager getBlock would have returned exception.
+ checkEvictedStatus(inputStream, 0, false);
+ }
+
+ /**
+ * The test expects AbfsInputStream to initiate a remote read request for
+ * the request offset and length when previous read ahead on the offset had failed.
+ * Also checks that the ReadBuffers are evicted as per the ReadBufferManager
+ * threshold criteria.
+ * @throws Exception
+ */
+ @Test
+ public void testOlderReadAheadFailure() throws Exception {
+ AbfsClient client = getMockAbfsClient();
+ AbfsRestOperation successOp = getMockRestOp();
+
+ // Stub :
+ // First Read request leads to 3 readahead calls: Fail all 3 readahead-client.read()
+ // A second read request will see that readahead had failed for data in
+ // the requested offset range and also that its is an older readahead request.
+ // So attempt a new read only for the requested range.
+ doThrow(new TimeoutException("Internal Server error for RAH-X"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Y"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Z"))
+ .doReturn(successOp) // pass the read for second read request
+ .doReturn(successOp) // pass success for post eviction test
+ .when(client)
+ .read(any(String.class), any(Long.class), any(byte[].class),
+ any(Integer.class), any(Integer.class), any(String.class));
+
+ AbfsInputStream inputStream = getAbfsInputStream(client, "testOlderReadAheadFailure.txt");
+
+ // First read request that fails as the readahead triggered from this request failed.
+ intercept(IOException.class,
+ () -> inputStream.read(new byte[1 * KILOBYTE]));
+
+ // Only the 3 readAhead threads should have triggered client.read
+ verifyReadCallCount(client, 3);
+
+ // Sleep for 30 sec so that the read ahead buffer qualifies for being old.
+ Thread.sleep(ReadBufferManager.getBufferManager().getThresholdAgeMilliseconds());
+
+ // Second read request should retry the read (and not issue any new readaheads)
+ inputStream.read(1 * KILOBYTE, new byte[1 * KILOBYTE], 0, 1 * KILOBYTE);
+
+ // Once created, mock will remember all interactions. So total number of read
+ // calls will be one more from earlier (there is a reset mock which will reset the
+ // count, but the mock stub is erased as well which needs AbsInputStream to be recreated,
+ // which beats the purpose)
+ verifyReadCallCount(client, 4);
+
+ // Stub returns success for the 5th read request, if ReadBuffers still
+ // persisted request would have failed for position 0.
+ checkEvictedStatus(inputStream, 0, false);
+ }
+
+ /**
+ * The test expects AbfsInputStream to utilize any data read ahead for
+ * requested offset and length.
+ * @throws Exception
+ */
+ @Test
+ public void testSuccessfulReadAhead() throws Exception {
+ // Mock failure for client.read()
+ AbfsClient client = getMockAbfsClient();
+
+ // Success operation mock
+ AbfsRestOperation op = getMockRestOp();
+
+ // Stub :
+ // Pass all readAheads and fail the post eviction request to
+ // prove ReadAhead buffer is used
+ // for post eviction check, fail all read aheads
+ doReturn(op)
+ .doReturn(op)
+ .doReturn(op)
+ .doThrow(new TimeoutException("Internal Server error for RAH-X"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Y"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Z"))
+ .when(client)
+ .read(any(String.class), any(Long.class), any(byte[].class),
+ any(Integer.class), any(Integer.class), any(String.class));
+
+ AbfsInputStream inputStream = getAbfsInputStream(client, "testSuccessfulReadAhead.txt");
+
+ // First read request that triggers readAheads.
+ inputStream.read(new byte[1 * KILOBYTE]);
+
+ // Only the 3 readAhead threads should have triggered client.read
+ verifyReadCallCount(client, 3);
+
+ // Another read request whose requested data is already read ahead.
+ inputStream.read(1 * KILOBYTE, new byte[1 * KILOBYTE], 0, 1 * KILOBYTE);
Review comment:
I'd go for replacing 1 * KILOBYTE with just KILOBYTE, or having constants _1KB, _2KB etc, which is what I've done elsewhere
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] hadoop-yetus commented on issue #1898: HADOOP-16852:
Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
hadoop-yetus commented on issue #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#issuecomment-604858193
:confetti_ball: **+1 overall**
| Vote | Subsystem | Runtime | Comment |
|:----:|----------:|--------:|:--------|
| +0 :ok: | reexec | 30m 18s | Docker mode activated. |
||| _ Prechecks _ |
| +1 :green_heart: | dupname | 0m 0s | No case conflicting files found. |
| +1 :green_heart: | @author | 0m 0s | The patch does not contain any @author tags. |
| +1 :green_heart: | test4tests | 0m 0s | The patch appears to include 1 new or modified test files. |
||| _ trunk Compile Tests _ |
| +1 :green_heart: | mvninstall | 22m 19s | trunk passed |
| +1 :green_heart: | compile | 0m 27s | trunk passed |
| +1 :green_heart: | checkstyle | 0m 20s | trunk passed |
| +1 :green_heart: | mvnsite | 0m 29s | trunk passed |
| +1 :green_heart: | shadedclient | 16m 15s | branch has no errors when building and testing our client artifacts. |
| +1 :green_heart: | javadoc | 0m 22s | trunk passed |
| +0 :ok: | spotbugs | 0m 51s | Used deprecated FindBugs config; considering switching to SpotBugs. |
| +1 :green_heart: | findbugs | 0m 48s | trunk passed |
| -0 :warning: | patch | 1m 6s | Used diff version of patch file. Binary files and potentially other changes not applied. Please rebase and squash commits if necessary. |
||| _ Patch Compile Tests _ |
| +1 :green_heart: | mvninstall | 0m 28s | the patch passed |
| +1 :green_heart: | compile | 0m 22s | the patch passed |
| +1 :green_heart: | javac | 0m 22s | the patch passed |
| -0 :warning: | checkstyle | 0m 15s | hadoop-tools/hadoop-azure: The patch generated 26 new + 0 unchanged - 0 fixed = 26 total (was 0) |
| +1 :green_heart: | mvnsite | 0m 25s | the patch passed |
| +1 :green_heart: | whitespace | 0m 0s | The patch has no whitespace issues. |
| +1 :green_heart: | shadedclient | 15m 50s | patch has no errors when building and testing our client artifacts. |
| +1 :green_heart: | javadoc | 0m 20s | the patch passed |
| +1 :green_heart: | findbugs | 0m 53s | the patch passed |
||| _ Other Tests _ |
| +1 :green_heart: | unit | 1m 22s | hadoop-azure in the patch passed. |
| +1 :green_heart: | asflicense | 0m 28s | The patch does not generate ASF License warnings. |
| | | 93m 2s | |
| Subsystem | Report/Notes |
|----------:|:-------------|
| Docker | Client=19.03.8 Server=19.03.8 base: https://builds.apache.org/job/hadoop-multibranch/job/PR-1898/4/artifact/out/Dockerfile |
| GITHUB PR | https://github.com/apache/hadoop/pull/1898 |
| JIRA Issue | HADOOP-16852 |
| Optional Tests | dupname asflicense compile javac javadoc mvninstall mvnsite unit shadedclient findbugs checkstyle |
| uname | Linux 4fa7f65cf8a9 4.15.0-74-generic #84-Ubuntu SMP Thu Dec 19 08:06:28 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux |
| Build tool | maven |
| Personality | personality/hadoop.sh |
| git revision | trunk / eaaaba1 |
| Default Java | 1.8.0_242 |
| checkstyle | https://builds.apache.org/job/hadoop-multibranch/job/PR-1898/4/artifact/out/diff-checkstyle-hadoop-tools_hadoop-azure.txt |
| Test Results | https://builds.apache.org/job/hadoop-multibranch/job/PR-1898/4/testReport/ |
| Max. process+thread count | 316 (vs. ulimit of 5500) |
| modules | C: hadoop-tools/hadoop-azure U: hadoop-tools/hadoop-azure |
| Console output | https://builds.apache.org/job/hadoop-multibranch/job/PR-1898/4/console |
| versions | git=2.7.4 maven=3.3.9 findbugs=3.1.0-RC1 |
| Powered by | Apache Yetus 0.11.1 https://yetus.apache.org |
This message was automatically generated.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] snvijaya commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
snvijaya commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r400113993
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBufferManager.java
##########
@@ -141,7 +144,8 @@ void queueReadAhead(final AbfsInputStream stream, final long requestedOffset, fi
* @param buffer the buffer to read data into. Note that the buffer will be written into from offset 0.
* @return the number of bytes read
*/
- int getBlock(final AbfsInputStream stream, final long position, final int length, final byte[] buffer) {
+ int getBlock(final AbfsInputStream stream, final long position, final int length, final byte[] buffer)
+ throws java.io.IOException {
Review comment:
Fixed
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] DadanielZ commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
DadanielZ commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r399725608
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBufferManager.java
##########
@@ -299,11 +327,32 @@ private void clearFromReadAheadQueue(final AbfsInputStream stream, final long re
}
private int getBlockFromCompletedQueue(final AbfsInputStream stream, final long position, final int length,
- final byte[] buffer) {
- ReadBuffer buf = getFromList(completedReadList, stream, position);
- if (buf == null || position >= buf.getOffset() + buf.getLength()) {
+ final byte[] buffer) throws IOException {
+ ReadBuffer buf = getBufferFromCompletedQueue(stream, position);
+
+ if (buf == null) {
return 0;
}
+
+ if (buf.getStatus() == ReadBufferStatus.READ_FAILED) {
+ // Eviction of a read buffer is triggered only when a queue request comes in
+ // and each eviction attempt tries to find one eligible buffer.
+ // Hence there are chances that an old read-ahead buffer with exception is still
+ // available. To prevent new read requests to fail due to such old buffers,
+ // return exception only from buffers that failed within last THRESHOLD_AGE_MILLISECONDS
+ if ((currentTimeMillis() - (buf.getTimeStamp()) < THRESHOLD_AGE_MILLISECONDS)) {
Review comment:
If this is to avoid the failure for new requests, then instead of checking the time window, why not reset the buffer status before throwing the exception? Then following new requests will not be affected by the old failure.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] snvijaya commented on issue #1898: HADOOP-16852: Report
read-ahead error back
Posted by GitBox <gi...@apache.org>.
snvijaya commented on issue #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#issuecomment-602409736
Made a fix where read-ahead thread will never read remote a length greater than its buffer size.
HNS enabled account:
[INFO] Tests run: 58, Failures: 0, Errors: 0, Skipped: 0
[WARNING] Tests run: 412, Failures: 0, Errors: 0, Skipped: 66
[WARNING] Tests run: 206, Failures: 0, Errors: 0, Skipped: 140
HNS not enabled account:
[INFO] Tests run: 58, Failures: 0, Errors: 0, Skipped: 0
[WARNING] Tests run: 412, Failures: 0, Errors: 0, Skipped: 240
[WARNING] Tests run: 206, Failures: 0, Errors: 0, Skipped: 140
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] DadanielZ commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
DadanielZ commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r399723723
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBuffer.java
##########
@@ -88,12 +92,23 @@ public void setBufferindex(int bufferindex) {
this.bufferindex = bufferindex;
}
+ public IOException getErrException() {
+ return errException;
+ }
+
+ public void setErrException(final java.io.IOException errException) {
Review comment:
IOException
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] steveloughran commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
steveloughran commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r405028733
##########
File path: hadoop-tools/hadoop-azure/src/test/java/org/apache/hadoop/fs/azurebfs/services/TestAbfsInputStream.java
##########
@@ -0,0 +1,433 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.hadoop.fs.azurebfs.services;
+
+import java.io.IOException;
+
+import org.junit.Assert;
+import org.junit.Test;
+
+import org.apache.hadoop.fs.azurebfs.AbstractAbfsIntegrationTest;
+import org.apache.hadoop.fs.azurebfs.contracts.exceptions.AzureBlobFileSystemException;
+import org.apache.hadoop.fs.azurebfs.contracts.exceptions.TimeoutException;
+
+import static org.mockito.ArgumentMatchers.any;
+import static org.mockito.Mockito.doReturn;
+import static org.mockito.Mockito.doThrow;
+import static org.mockito.Mockito.mock;
+import static org.mockito.Mockito.times;
+import static org.mockito.Mockito.verify;
+import static org.mockito.Mockito.when;
+import static org.apache.hadoop.test.LambdaTestUtils.intercept;
+import static org.apache.hadoop.fs.azurebfs.constants.AbfsHttpConstants.FORWARD_SLASH;
+
+/**
+ * Unit test AbfsInputStream.
+ */
+public class TestAbfsInputStream extends
+ AbstractAbfsIntegrationTest {
+
+ private static final int KILOBYTE = 1024;
+
+ private AbfsRestOperation getMockRestOp() {
+ AbfsRestOperation op = mock(AbfsRestOperation.class);
+ AbfsHttpOperation httpOp = mock(AbfsHttpOperation.class);
+ when(httpOp.getBytesReceived()).thenReturn(1024L);
+ when(op.getResult()).thenReturn(httpOp);
+ return op;
+ }
+
+ private AbfsClient getMockAbfsClient() {
+ // Mock failure for client.read()
+ AbfsClient client = mock(AbfsClient.class);
+ AbfsPerfTracker tracker = new AbfsPerfTracker(
+ "test",
+ this.getAccountName(),
+ this.getConfiguration());
+ when(client.getAbfsPerfTracker()).thenReturn(tracker);
+
+ return client;
+ }
+
+ private AbfsInputStream getAbfsInputStream(AbfsClient mockAbfsClient, String fileName) {
+ // Create AbfsInputStream with the client instance
+ AbfsInputStream inputStream = new AbfsInputStream(
+ mockAbfsClient,
+ null,
+ FORWARD_SLASH + fileName,
+ 3 * KILOBYTE,
+ 1 * KILOBYTE, // Setting read ahead buffer size of 1 KB
+ this.getConfiguration().getReadAheadQueueDepth(),
+ this.getConfiguration().getTolerateOobAppends(),
+ "eTag");
+
+ return inputStream;
+ }
+
+ private void queueReadAheads(AbfsInputStream inputStream) {
+ // Mimic AbfsInputStream readAhead queue requests
+ ReadBufferManager.getBufferManager()
+ .queueReadAhead(inputStream, 0, 1 * KILOBYTE);
+ ReadBufferManager.getBufferManager()
+ .queueReadAhead(inputStream, 1 * KILOBYTE, 1 * KILOBYTE);
+ ReadBufferManager.getBufferManager()
+ .queueReadAhead(inputStream, 2 * KILOBYTE, 1 * KILOBYTE);
+ }
+
+ private void verifyReadCallCount(AbfsClient client, int count) throws
+ AzureBlobFileSystemException, InterruptedException {
+ // ReadAhead threads are triggered asynchronously.
+ // Wait a second before verifying the number of total calls.
+ Thread.sleep(1000);
+ verify(client, times(count)).read(any(String.class), any(Long.class),
+ any(byte[].class), any(Integer.class), any(Integer.class),
+ any(String.class));
+ }
+
+ private void checkEvictedStatus(AbfsInputStream inputStream, int position, boolean expectedToThrowException)
+ throws Exception {
+ // Sleep for the eviction threshold time
+ Thread.sleep(ReadBufferManager.getBufferManager().getThresholdAgeMilliseconds() + 1000);
+
+ // Eviction is done only when AbfsInputStream tries to queue new items.
+ // 1 tryEvict will remove 1 eligible item. To ensure that the current test buffer
+ // will get evicted (considering there could be other tests running in parallel),
+ // call tryEvict for the number of items that are there in completedReadList.
+ int numOfCompletedReadListItems = ReadBufferManager.getBufferManager().getCompletedReadListSize();
+ while (numOfCompletedReadListItems > 0) {
+ ReadBufferManager.getBufferManager().callTryEvict();
+ numOfCompletedReadListItems--;
+ }
+
+ if (expectedToThrowException) {
+ intercept(IOException.class,
+ () -> inputStream.read(position, new byte[1 * KILOBYTE], 0, 1 * KILOBYTE));
+ } else {
+ inputStream.read(position, new byte[1 * KILOBYTE], 0, 1 * KILOBYTE);
+ }
+ }
+
+ public TestAbfsInputStream() throws Exception {
+ super();
+ }
+
+ /**
+ * This test expects AbfsInputStream to throw the exception that readAhead
+ * thread received on read. The readAhead thread must be initiated from the
+ * active read request itself.
+ * Also checks that the ReadBuffers are evicted as per the ReadBufferManager
+ * threshold criteria.
+ * @throws Exception
+ */
+ @Test
+ public void testFailedReadAhead() throws Exception {
+ AbfsClient client = getMockAbfsClient();
+ AbfsRestOperation successOp = getMockRestOp();
+
+ // Stub :
+ // Read request leads to 3 readahead calls: Fail all 3 readahead-client.read()
+ // Actual read request fails with the failure in readahead thread
+ doThrow(new TimeoutException("Internal Server error for RAH-Thread-X"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Thread-Y"))
+ .doThrow(new TimeoutException("Internal Server error RAH-Thread-Z"))
+ .doReturn(successOp) // Any extra calls to read, pass it.
+ .when(client)
+ .read(any(String.class), any(Long.class), any(byte[].class),
+ any(Integer.class), any(Integer.class), any(String.class));
+
+ AbfsInputStream inputStream = getAbfsInputStream(client, "testFailedReadAhead.txt");
+
+ // Scenario: ReadAhead triggered from current active read call failed
+ // Before the change to return exception from readahead buffer,
+ // AbfsInputStream would have triggered an extra readremote on noticing
+ // data absent in readahead buffers
+ // In this test, a read should trigger 3 client.read() calls as file is 3 KB
+ // and readahead buffer size set in AbfsInputStream is 1 KB
+ // There should only be a total of 3 client.read() in this test.
+ intercept(IOException.class,
+ () -> inputStream.read(new byte[1 * KILOBYTE]));
+
+ // Only the 3 readAhead threads should have triggered client.read
+ verifyReadCallCount(client, 3);
+
+ // Stub returns success for the 4th read request, if ReadBuffers still
+ // persisted, ReadAheadManager getBlock would have returned exception.
+ checkEvictedStatus(inputStream, 0, false);
+ }
+
+ /**
+ * The test expects AbfsInputStream to initiate a remote read request for
+ * the request offset and length when previous read ahead on the offset had failed.
+ * Also checks that the ReadBuffers are evicted as per the ReadBufferManager
+ * threshold criteria.
+ * @throws Exception
+ */
+ @Test
+ public void testOlderReadAheadFailure() throws Exception {
+ AbfsClient client = getMockAbfsClient();
+ AbfsRestOperation successOp = getMockRestOp();
+
+ // Stub :
+ // First Read request leads to 3 readahead calls: Fail all 3 readahead-client.read()
+ // A second read request will see that readahead had failed for data in
+ // the requested offset range and also that its is an older readahead request.
+ // So attempt a new read only for the requested range.
+ doThrow(new TimeoutException("Internal Server error for RAH-X"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Y"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Z"))
+ .doReturn(successOp) // pass the read for second read request
+ .doReturn(successOp) // pass success for post eviction test
+ .when(client)
+ .read(any(String.class), any(Long.class), any(byte[].class),
+ any(Integer.class), any(Integer.class), any(String.class));
+
+ AbfsInputStream inputStream = getAbfsInputStream(client, "testOlderReadAheadFailure.txt");
+
+ // First read request that fails as the readahead triggered from this request failed.
+ intercept(IOException.class,
+ () -> inputStream.read(new byte[1 * KILOBYTE]));
+
+ // Only the 3 readAhead threads should have triggered client.read
+ verifyReadCallCount(client, 3);
+
+ // Sleep for 30 sec so that the read ahead buffer qualifies for being old.
+ Thread.sleep(ReadBufferManager.getBufferManager().getThresholdAgeMilliseconds());
+
+ // Second read request should retry the read (and not issue any new readaheads)
+ inputStream.read(1 * KILOBYTE, new byte[1 * KILOBYTE], 0, 1 * KILOBYTE);
+
+ // Once created, mock will remember all interactions. So total number of read
+ // calls will be one more from earlier (there is a reset mock which will reset the
+ // count, but the mock stub is erased as well which needs AbsInputStream to be recreated,
+ // which beats the purpose)
+ verifyReadCallCount(client, 4);
+
+ // Stub returns success for the 5th read request, if ReadBuffers still
+ // persisted request would have failed for position 0.
+ checkEvictedStatus(inputStream, 0, false);
+ }
+
+ /**
+ * The test expects AbfsInputStream to utilize any data read ahead for
+ * requested offset and length.
+ * @throws Exception
+ */
+ @Test
+ public void testSuccessfulReadAhead() throws Exception {
+ // Mock failure for client.read()
+ AbfsClient client = getMockAbfsClient();
+
+ // Success operation mock
+ AbfsRestOperation op = getMockRestOp();
+
+ // Stub :
+ // Pass all readAheads and fail the post eviction request to
+ // prove ReadAhead buffer is used
+ // for post eviction check, fail all read aheads
+ doReturn(op)
+ .doReturn(op)
+ .doReturn(op)
+ .doThrow(new TimeoutException("Internal Server error for RAH-X"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Y"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Z"))
+ .when(client)
+ .read(any(String.class), any(Long.class), any(byte[].class),
+ any(Integer.class), any(Integer.class), any(String.class));
+
+ AbfsInputStream inputStream = getAbfsInputStream(client, "testSuccessfulReadAhead.txt");
+
+ // First read request that triggers readAheads.
+ inputStream.read(new byte[1 * KILOBYTE]);
+
+ // Only the 3 readAhead threads should have triggered client.read
+ verifyReadCallCount(client, 3);
+
+ // Another read request whose requested data is already read ahead.
+ inputStream.read(1 * KILOBYTE, new byte[1 * KILOBYTE], 0, 1 * KILOBYTE);
+
+ // Once created, mock will remember all interactions.
+ // As the above read should not have triggered any server calls, total
+ // number of read calls made at this point will be same as last.
+ verifyReadCallCount(client, 3);
+
+ // Stub will throw exception for client.read() for 4th and later calls
+ // if not using the read-ahead buffer exception will be thrown on read
+ checkEvictedStatus(inputStream, 0, true);
+ }
+
+ /**
+ * This test expects ReadAheadManager to throw exception if the read ahead
+ * thread had failed within the last 30 sec.
+ * Also checks that the ReadBuffers are evicted as per the ReadBufferManager
+ * threshold criteria.
+ * @throws Exception
+ */
+ @Test
+ public void testReadAheadManagerForFailedReadAhead() throws Exception {
+ AbfsClient client = getMockAbfsClient();
+ AbfsRestOperation successOp = getMockRestOp();
+
+ // Stub :
+ // Read request leads to 3 readahead calls: Fail all 3 readahead-client.read()
+ // Actual read request fails with the failure in readahead thread
+ doThrow(new TimeoutException("Internal Server error for RAH-Thread-X"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Thread-Y"))
+ .doThrow(new TimeoutException("Internal Server error RAH-Thread-Z"))
+ .doReturn(successOp) // Any extra calls to read, pass it.
+ .when(client)
+ .read(any(String.class), any(Long.class), any(byte[].class),
+ any(Integer.class), any(Integer.class), any(String.class));
+
+ AbfsInputStream inputStream = getAbfsInputStream(client, "testReadAheadManagerForFailedReadAhead.txt");
+
+ queueReadAheads(inputStream);
+
+ // AbfsInputStream Read would have waited for the read-ahead for the requested offset
+ // as we are testing from ReadAheadManager directly, sleep for a sec to
+ // get the read ahead threads to complete
+ Thread.sleep(1000);
+
+ // if readAhead failed for specific offset, getBlock should
+ // throw exception from the ReadBuffer that failed within last 30 sec
+ intercept(IOException.class,
+ () -> ReadBufferManager.getBufferManager().getBlock(
+ inputStream,
+ 0,
+ 1 * KILOBYTE,
+ new byte[1 * KILOBYTE]));
+
+ // Only the 3 readAhead threads should have triggered client.read
+ verifyReadCallCount(client, 3);
+
+ // Stub returns success for the 4th read request, if ReadBuffers still
+ // persisted, ReadAheadManager getBlock would have returned exception.
+ checkEvictedStatus(inputStream, 0, false);
+ }
+
+ /**
+ * The test expects ReadAheadManager to return 0 receivedBytes when previous
+ * read ahead on the offset had failed and not throw exception received then.
+ * Also checks that the ReadBuffers are evicted as per the ReadBufferManager
+ * threshold criteria.
+ * @throws Exception
+ */
+ @Test
+ public void testReadAheadManagerForOlderReadAheadFailure() throws Exception {
+ AbfsClient client = getMockAbfsClient();
+ AbfsRestOperation successOp = getMockRestOp();
+
+ // Stub :
+ // First Read request leads to 3 readahead calls: Fail all 3 readahead-client.read()
+ // A second read request will see that readahead had failed for data in
+ // the requested offset range but also that its is an older readahead request.
+ // System issue could have resolved by now, so attempt a new read only for the requested range.
+ doThrow(new TimeoutException("Internal Server error for RAH-X"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-X"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-X"))
+ .doReturn(successOp) // pass the read for second read request
+ .doReturn(successOp) // pass success for post eviction test
+ .when(client)
+ .read(any(String.class), any(Long.class), any(byte[].class),
+ any(Integer.class), any(Integer.class), any(String.class));
+
+ AbfsInputStream inputStream = getAbfsInputStream(client, "testReadAheadManagerForOlderReadAheadFailure.txt");
+
+ queueReadAheads(inputStream);
+
+ // AbfsInputStream Read would have waited for the read-ahead for the requested offset
+ // as we are testing from ReadAheadManager directly, sleep for 30 secs so that
+ // read buffer qualifies for to be an old buffer
+ Thread.sleep(ReadBufferManager.getBufferManager().getThresholdAgeMilliseconds());
Review comment:
is there any way for test runs to avoid these long sleeps? This might add 30s to the test run, and that adds up over the day.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] steveloughran commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
steveloughran commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r405023982
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBuffer.java
##########
@@ -17,11 +17,13 @@
*/
package org.apache.hadoop.fs.azurebfs.services;
-
+import java.io.IOException;
Review comment:
add space above this line
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] snvijaya commented on issue #1898: HADOOP-16852: Report
read-ahead error back
Posted by GitBox <gi...@apache.org>.
snvijaya commented on issue #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#issuecomment-601116210
Test results:
HNS enabled account:
[INFO] Tests run: 58, Failures: 0, Errors: 0, Skipped: 0
[WARNING] Tests run: 412, Failures: 0, Errors: 0, Skipped: 66
[WARNING] Tests run: 206, Failures: 0, Errors: 0, Skipped: 140
HNS not enabled account:
[INFO] Tests run: 58, Failures: 0, Errors: 0, Skipped: 0
[WARNING] Tests run: 412, Failures: 0, Errors: 0, Skipped: 240
[WARNING] Tests run: 206, Failures: 0, Errors: 0, Skipped: 140
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] goiri commented on a change in pull request #1898: Report
read-ahead error back
Posted by GitBox <gi...@apache.org>.
goiri commented on a change in pull request #1898: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r393792745
##########
File path: hadoop-tools/hadoop-azure/src/test/java/org/apache/hadoop/fs/azurebfs/services/TestAbfsInputStream.java
##########
@@ -0,0 +1,438 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.hadoop.fs.azurebfs.services;
+
+import java.io.IOException;
+import java.util.UUID;
+
+import org.junit.Assert;
+import org.junit.Test;
+
+import org.apache.hadoop.fs.azurebfs.AbstractAbfsIntegrationTest;
+import org.apache.hadoop.fs.azurebfs.contracts.exceptions.AzureBlobFileSystemException;
+import org.apache.hadoop.fs.azurebfs.contracts.exceptions.TimeoutException;
+
+import static java.util.UUID.randomUUID;
+import static org.mockito.ArgumentMatchers.any;
+import static org.mockito.Mockito.doReturn;
+import static org.mockito.Mockito.doThrow;
+import static org.mockito.Mockito.mock;
+import static org.mockito.Mockito.times;
+import static org.mockito.Mockito.verify;
+import static org.mockito.Mockito.when;
+import static org.apache.hadoop.test.LambdaTestUtils.intercept;
+import static org.apache.hadoop.fs.azurebfs.constants.AbfsHttpConstants.FORWARD_SLASH;
+
+/**
+ * Unit test AbfsInputStream.
+ */
+public class TestAbfsInputStream extends
Review comment:
What about capturing logs and checking for the messages?
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] snvijaya commented on issue #1898: HADOOP-16852: Report
read-ahead error back
Posted by GitBox <gi...@apache.org>.
snvijaya commented on issue #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#issuecomment-605941359
@DadanielZ - Thanks for the review. I have left the comment on the bufferstatus versus the timestamp check unresolved. As mentioned in my comments, the intention is to throw the exception from the read-ahead buffer for any reads that qualify the buffer's offset and length range. Please let me know if you have any concerns.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] snvijaya commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
snvijaya commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r394942426
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBufferManager.java
##########
@@ -101,6 +107,7 @@ void queueReadAhead(final AbfsInputStream stream, final long requestedOffset, fi
if (isAlreadyQueued(stream, requestedOffset)) {
return; // already queued, do not queue again
}
+
Review comment:
Done
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] snvijaya commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
snvijaya commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r394942253
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBufferManager.java
##########
@@ -141,7 +149,8 @@ void queueReadAhead(final AbfsInputStream stream, final long requestedOffset, fi
* @param buffer the buffer to read data into. Note that the buffer will be written into from offset 0.
* @return the number of bytes read
Review comment:
New param undone as part of latest iteration.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] goiri commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
goiri commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r395128849
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBufferManager.java
##########
@@ -299,11 +327,32 @@ private void clearFromReadAheadQueue(final AbfsInputStream stream, final long re
}
private int getBlockFromCompletedQueue(final AbfsInputStream stream, final long position, final int length,
- final byte[] buffer) {
- ReadBuffer buf = getFromList(completedReadList, stream, position);
- if (buf == null || position >= buf.getOffset() + buf.getLength()) {
+ final byte[] buffer) throws IOException {
+ ReadBuffer buf = getBufferFromCompletedQueue(stream, position);
+
+ if (buf == null) {
return 0;
}
+
+ if (buf.getStatus() == ReadBufferStatus.READ_FAILED) {
+ // Eviction of a read buffer is triggered only when a queue request comes in
+ // and each eviction attempt tries to find one eligible buffer.
+ // Hence there are chances that an old read-ahead buffer with exception is still
+ // available. To prevent new read requests to fail due to such old buffers,
+ // return exception only from buffers that failed within last THRESHOLD_AGE_MILLISECONDS
+ if ((currentTimeMillis() - (buf.getTimeStamp()) < THRESHOLD_AGE_MILLISECONDS)) {
Review comment:
These parenthesis are confusing.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] steveloughran commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
steveloughran commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r405033372
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBufferManager.java
##########
@@ -299,11 +327,32 @@ private void clearFromReadAheadQueue(final AbfsInputStream stream, final long re
}
private int getBlockFromCompletedQueue(final AbfsInputStream stream, final long position, final int length,
- final byte[] buffer) {
- ReadBuffer buf = getFromList(completedReadList, stream, position);
- if (buf == null || position >= buf.getOffset() + buf.getLength()) {
+ final byte[] buffer) throws IOException {
+ ReadBuffer buf = getBufferFromCompletedQueue(stream, position);
+
+ if (buf == null) {
return 0;
}
+
+ if (buf.getStatus() == ReadBufferStatus.READ_FAILED) {
+ // Eviction of a read buffer is triggered only when a queue request comes in
+ // and each eviction attempt tries to find one eligible buffer.
+ // Hence there are chances that an old read-ahead buffer with exception is still
+ // available. To prevent new read requests to fail due to such old buffers,
+ // return exception only from buffers that failed within last THRESHOLD_AGE_MILLISECONDS
+ if ((currentTimeMillis() - (buf.getTimeStamp()) < THRESHOLD_AGE_MILLISECONDS)) {
Review comment:
why 30s? Any way to tune this for test runs?
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] steveloughran commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
steveloughran commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r405027481
##########
File path: hadoop-tools/hadoop-azure/src/test/java/org/apache/hadoop/fs/azurebfs/services/TestAbfsInputStream.java
##########
@@ -0,0 +1,433 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.hadoop.fs.azurebfs.services;
+
+import java.io.IOException;
+
+import org.junit.Assert;
+import org.junit.Test;
+
+import org.apache.hadoop.fs.azurebfs.AbstractAbfsIntegrationTest;
+import org.apache.hadoop.fs.azurebfs.contracts.exceptions.AzureBlobFileSystemException;
+import org.apache.hadoop.fs.azurebfs.contracts.exceptions.TimeoutException;
+
+import static org.mockito.ArgumentMatchers.any;
+import static org.mockito.Mockito.doReturn;
+import static org.mockito.Mockito.doThrow;
+import static org.mockito.Mockito.mock;
+import static org.mockito.Mockito.times;
+import static org.mockito.Mockito.verify;
+import static org.mockito.Mockito.when;
+import static org.apache.hadoop.test.LambdaTestUtils.intercept;
+import static org.apache.hadoop.fs.azurebfs.constants.AbfsHttpConstants.FORWARD_SLASH;
+
+/**
+ * Unit test AbfsInputStream.
+ */
+public class TestAbfsInputStream extends
+ AbstractAbfsIntegrationTest {
+
+ private static final int KILOBYTE = 1024;
+
+ private AbfsRestOperation getMockRestOp() {
+ AbfsRestOperation op = mock(AbfsRestOperation.class);
+ AbfsHttpOperation httpOp = mock(AbfsHttpOperation.class);
+ when(httpOp.getBytesReceived()).thenReturn(1024L);
+ when(op.getResult()).thenReturn(httpOp);
+ return op;
+ }
+
+ private AbfsClient getMockAbfsClient() {
+ // Mock failure for client.read()
+ AbfsClient client = mock(AbfsClient.class);
+ AbfsPerfTracker tracker = new AbfsPerfTracker(
+ "test",
+ this.getAccountName(),
+ this.getConfiguration());
+ when(client.getAbfsPerfTracker()).thenReturn(tracker);
+
+ return client;
+ }
+
+ private AbfsInputStream getAbfsInputStream(AbfsClient mockAbfsClient, String fileName) {
+ // Create AbfsInputStream with the client instance
+ AbfsInputStream inputStream = new AbfsInputStream(
+ mockAbfsClient,
+ null,
+ FORWARD_SLASH + fileName,
+ 3 * KILOBYTE,
+ 1 * KILOBYTE, // Setting read ahead buffer size of 1 KB
+ this.getConfiguration().getReadAheadQueueDepth(),
+ this.getConfiguration().getTolerateOobAppends(),
+ "eTag");
+
+ return inputStream;
+ }
+
+ private void queueReadAheads(AbfsInputStream inputStream) {
+ // Mimic AbfsInputStream readAhead queue requests
+ ReadBufferManager.getBufferManager()
+ .queueReadAhead(inputStream, 0, 1 * KILOBYTE);
+ ReadBufferManager.getBufferManager()
+ .queueReadAhead(inputStream, 1 * KILOBYTE, 1 * KILOBYTE);
+ ReadBufferManager.getBufferManager()
+ .queueReadAhead(inputStream, 2 * KILOBYTE, 1 * KILOBYTE);
+ }
+
+ private void verifyReadCallCount(AbfsClient client, int count) throws
+ AzureBlobFileSystemException, InterruptedException {
+ // ReadAhead threads are triggered asynchronously.
+ // Wait a second before verifying the number of total calls.
+ Thread.sleep(1000);
+ verify(client, times(count)).read(any(String.class), any(Long.class),
+ any(byte[].class), any(Integer.class), any(Integer.class),
+ any(String.class));
+ }
+
+ private void checkEvictedStatus(AbfsInputStream inputStream, int position, boolean expectedToThrowException)
+ throws Exception {
+ // Sleep for the eviction threshold time
+ Thread.sleep(ReadBufferManager.getBufferManager().getThresholdAgeMilliseconds() + 1000);
+
+ // Eviction is done only when AbfsInputStream tries to queue new items.
+ // 1 tryEvict will remove 1 eligible item. To ensure that the current test buffer
+ // will get evicted (considering there could be other tests running in parallel),
+ // call tryEvict for the number of items that are there in completedReadList.
+ int numOfCompletedReadListItems = ReadBufferManager.getBufferManager().getCompletedReadListSize();
+ while (numOfCompletedReadListItems > 0) {
+ ReadBufferManager.getBufferManager().callTryEvict();
+ numOfCompletedReadListItems--;
+ }
+
+ if (expectedToThrowException) {
+ intercept(IOException.class,
+ () -> inputStream.read(position, new byte[1 * KILOBYTE], 0, 1 * KILOBYTE));
+ } else {
+ inputStream.read(position, new byte[1 * KILOBYTE], 0, 1 * KILOBYTE);
+ }
+ }
+
+ public TestAbfsInputStream() throws Exception {
Review comment:
not needed; just cut it
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] steveloughran commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
steveloughran commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r405028966
##########
File path: hadoop-tools/hadoop-azure/src/test/java/org/apache/hadoop/fs/azurebfs/services/TestAbfsInputStream.java
##########
@@ -0,0 +1,433 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.hadoop.fs.azurebfs.services;
+
+import java.io.IOException;
+
+import org.junit.Assert;
+import org.junit.Test;
+
+import org.apache.hadoop.fs.azurebfs.AbstractAbfsIntegrationTest;
+import org.apache.hadoop.fs.azurebfs.contracts.exceptions.AzureBlobFileSystemException;
+import org.apache.hadoop.fs.azurebfs.contracts.exceptions.TimeoutException;
+
+import static org.mockito.ArgumentMatchers.any;
+import static org.mockito.Mockito.doReturn;
+import static org.mockito.Mockito.doThrow;
+import static org.mockito.Mockito.mock;
+import static org.mockito.Mockito.times;
+import static org.mockito.Mockito.verify;
+import static org.mockito.Mockito.when;
+import static org.apache.hadoop.test.LambdaTestUtils.intercept;
+import static org.apache.hadoop.fs.azurebfs.constants.AbfsHttpConstants.FORWARD_SLASH;
+
+/**
+ * Unit test AbfsInputStream.
+ */
+public class TestAbfsInputStream extends
+ AbstractAbfsIntegrationTest {
+
+ private static final int KILOBYTE = 1024;
+
+ private AbfsRestOperation getMockRestOp() {
+ AbfsRestOperation op = mock(AbfsRestOperation.class);
+ AbfsHttpOperation httpOp = mock(AbfsHttpOperation.class);
+ when(httpOp.getBytesReceived()).thenReturn(1024L);
+ when(op.getResult()).thenReturn(httpOp);
+ return op;
+ }
+
+ private AbfsClient getMockAbfsClient() {
+ // Mock failure for client.read()
+ AbfsClient client = mock(AbfsClient.class);
+ AbfsPerfTracker tracker = new AbfsPerfTracker(
+ "test",
+ this.getAccountName(),
+ this.getConfiguration());
+ when(client.getAbfsPerfTracker()).thenReturn(tracker);
+
+ return client;
+ }
+
+ private AbfsInputStream getAbfsInputStream(AbfsClient mockAbfsClient, String fileName) {
+ // Create AbfsInputStream with the client instance
+ AbfsInputStream inputStream = new AbfsInputStream(
+ mockAbfsClient,
+ null,
+ FORWARD_SLASH + fileName,
+ 3 * KILOBYTE,
+ 1 * KILOBYTE, // Setting read ahead buffer size of 1 KB
+ this.getConfiguration().getReadAheadQueueDepth(),
+ this.getConfiguration().getTolerateOobAppends(),
+ "eTag");
+
+ return inputStream;
+ }
+
+ private void queueReadAheads(AbfsInputStream inputStream) {
+ // Mimic AbfsInputStream readAhead queue requests
+ ReadBufferManager.getBufferManager()
+ .queueReadAhead(inputStream, 0, 1 * KILOBYTE);
+ ReadBufferManager.getBufferManager()
+ .queueReadAhead(inputStream, 1 * KILOBYTE, 1 * KILOBYTE);
+ ReadBufferManager.getBufferManager()
+ .queueReadAhead(inputStream, 2 * KILOBYTE, 1 * KILOBYTE);
+ }
+
+ private void verifyReadCallCount(AbfsClient client, int count) throws
+ AzureBlobFileSystemException, InterruptedException {
+ // ReadAhead threads are triggered asynchronously.
+ // Wait a second before verifying the number of total calls.
+ Thread.sleep(1000);
+ verify(client, times(count)).read(any(String.class), any(Long.class),
+ any(byte[].class), any(Integer.class), any(Integer.class),
+ any(String.class));
+ }
+
+ private void checkEvictedStatus(AbfsInputStream inputStream, int position, boolean expectedToThrowException)
+ throws Exception {
+ // Sleep for the eviction threshold time
+ Thread.sleep(ReadBufferManager.getBufferManager().getThresholdAgeMilliseconds() + 1000);
+
+ // Eviction is done only when AbfsInputStream tries to queue new items.
+ // 1 tryEvict will remove 1 eligible item. To ensure that the current test buffer
+ // will get evicted (considering there could be other tests running in parallel),
+ // call tryEvict for the number of items that are there in completedReadList.
+ int numOfCompletedReadListItems = ReadBufferManager.getBufferManager().getCompletedReadListSize();
+ while (numOfCompletedReadListItems > 0) {
+ ReadBufferManager.getBufferManager().callTryEvict();
+ numOfCompletedReadListItems--;
+ }
+
+ if (expectedToThrowException) {
+ intercept(IOException.class,
+ () -> inputStream.read(position, new byte[1 * KILOBYTE], 0, 1 * KILOBYTE));
+ } else {
+ inputStream.read(position, new byte[1 * KILOBYTE], 0, 1 * KILOBYTE);
+ }
+ }
+
+ public TestAbfsInputStream() throws Exception {
+ super();
+ }
+
+ /**
+ * This test expects AbfsInputStream to throw the exception that readAhead
+ * thread received on read. The readAhead thread must be initiated from the
+ * active read request itself.
+ * Also checks that the ReadBuffers are evicted as per the ReadBufferManager
+ * threshold criteria.
+ * @throws Exception
+ */
+ @Test
+ public void testFailedReadAhead() throws Exception {
+ AbfsClient client = getMockAbfsClient();
+ AbfsRestOperation successOp = getMockRestOp();
+
+ // Stub :
+ // Read request leads to 3 readahead calls: Fail all 3 readahead-client.read()
+ // Actual read request fails with the failure in readahead thread
+ doThrow(new TimeoutException("Internal Server error for RAH-Thread-X"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Thread-Y"))
+ .doThrow(new TimeoutException("Internal Server error RAH-Thread-Z"))
+ .doReturn(successOp) // Any extra calls to read, pass it.
+ .when(client)
+ .read(any(String.class), any(Long.class), any(byte[].class),
+ any(Integer.class), any(Integer.class), any(String.class));
+
+ AbfsInputStream inputStream = getAbfsInputStream(client, "testFailedReadAhead.txt");
+
+ // Scenario: ReadAhead triggered from current active read call failed
+ // Before the change to return exception from readahead buffer,
+ // AbfsInputStream would have triggered an extra readremote on noticing
+ // data absent in readahead buffers
+ // In this test, a read should trigger 3 client.read() calls as file is 3 KB
+ // and readahead buffer size set in AbfsInputStream is 1 KB
+ // There should only be a total of 3 client.read() in this test.
+ intercept(IOException.class,
+ () -> inputStream.read(new byte[1 * KILOBYTE]));
+
+ // Only the 3 readAhead threads should have triggered client.read
+ verifyReadCallCount(client, 3);
+
+ // Stub returns success for the 4th read request, if ReadBuffers still
+ // persisted, ReadAheadManager getBlock would have returned exception.
+ checkEvictedStatus(inputStream, 0, false);
+ }
+
+ /**
+ * The test expects AbfsInputStream to initiate a remote read request for
+ * the request offset and length when previous read ahead on the offset had failed.
+ * Also checks that the ReadBuffers are evicted as per the ReadBufferManager
+ * threshold criteria.
+ * @throws Exception
+ */
+ @Test
+ public void testOlderReadAheadFailure() throws Exception {
+ AbfsClient client = getMockAbfsClient();
+ AbfsRestOperation successOp = getMockRestOp();
+
+ // Stub :
+ // First Read request leads to 3 readahead calls: Fail all 3 readahead-client.read()
+ // A second read request will see that readahead had failed for data in
+ // the requested offset range and also that its is an older readahead request.
+ // So attempt a new read only for the requested range.
+ doThrow(new TimeoutException("Internal Server error for RAH-X"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Y"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Z"))
+ .doReturn(successOp) // pass the read for second read request
+ .doReturn(successOp) // pass success for post eviction test
+ .when(client)
+ .read(any(String.class), any(Long.class), any(byte[].class),
+ any(Integer.class), any(Integer.class), any(String.class));
+
+ AbfsInputStream inputStream = getAbfsInputStream(client, "testOlderReadAheadFailure.txt");
+
+ // First read request that fails as the readahead triggered from this request failed.
+ intercept(IOException.class,
+ () -> inputStream.read(new byte[1 * KILOBYTE]));
+
+ // Only the 3 readAhead threads should have triggered client.read
+ verifyReadCallCount(client, 3);
+
+ // Sleep for 30 sec so that the read ahead buffer qualifies for being old.
+ Thread.sleep(ReadBufferManager.getBufferManager().getThresholdAgeMilliseconds());
+
+ // Second read request should retry the read (and not issue any new readaheads)
+ inputStream.read(1 * KILOBYTE, new byte[1 * KILOBYTE], 0, 1 * KILOBYTE);
+
+ // Once created, mock will remember all interactions. So total number of read
+ // calls will be one more from earlier (there is a reset mock which will reset the
+ // count, but the mock stub is erased as well which needs AbsInputStream to be recreated,
+ // which beats the purpose)
+ verifyReadCallCount(client, 4);
+
+ // Stub returns success for the 5th read request, if ReadBuffers still
+ // persisted request would have failed for position 0.
+ checkEvictedStatus(inputStream, 0, false);
+ }
+
+ /**
+ * The test expects AbfsInputStream to utilize any data read ahead for
+ * requested offset and length.
+ * @throws Exception
+ */
+ @Test
+ public void testSuccessfulReadAhead() throws Exception {
+ // Mock failure for client.read()
+ AbfsClient client = getMockAbfsClient();
+
+ // Success operation mock
+ AbfsRestOperation op = getMockRestOp();
+
+ // Stub :
+ // Pass all readAheads and fail the post eviction request to
+ // prove ReadAhead buffer is used
+ // for post eviction check, fail all read aheads
+ doReturn(op)
+ .doReturn(op)
+ .doReturn(op)
+ .doThrow(new TimeoutException("Internal Server error for RAH-X"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Y"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Z"))
+ .when(client)
+ .read(any(String.class), any(Long.class), any(byte[].class),
+ any(Integer.class), any(Integer.class), any(String.class));
+
+ AbfsInputStream inputStream = getAbfsInputStream(client, "testSuccessfulReadAhead.txt");
+
+ // First read request that triggers readAheads.
+ inputStream.read(new byte[1 * KILOBYTE]);
+
+ // Only the 3 readAhead threads should have triggered client.read
+ verifyReadCallCount(client, 3);
+
+ // Another read request whose requested data is already read ahead.
+ inputStream.read(1 * KILOBYTE, new byte[1 * KILOBYTE], 0, 1 * KILOBYTE);
+
+ // Once created, mock will remember all interactions.
+ // As the above read should not have triggered any server calls, total
+ // number of read calls made at this point will be same as last.
+ verifyReadCallCount(client, 3);
+
+ // Stub will throw exception for client.read() for 4th and later calls
+ // if not using the read-ahead buffer exception will be thrown on read
+ checkEvictedStatus(inputStream, 0, true);
+ }
+
+ /**
+ * This test expects ReadAheadManager to throw exception if the read ahead
+ * thread had failed within the last 30 sec.
+ * Also checks that the ReadBuffers are evicted as per the ReadBufferManager
+ * threshold criteria.
+ * @throws Exception
+ */
+ @Test
+ public void testReadAheadManagerForFailedReadAhead() throws Exception {
+ AbfsClient client = getMockAbfsClient();
+ AbfsRestOperation successOp = getMockRestOp();
+
+ // Stub :
+ // Read request leads to 3 readahead calls: Fail all 3 readahead-client.read()
+ // Actual read request fails with the failure in readahead thread
+ doThrow(new TimeoutException("Internal Server error for RAH-Thread-X"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-Thread-Y"))
+ .doThrow(new TimeoutException("Internal Server error RAH-Thread-Z"))
+ .doReturn(successOp) // Any extra calls to read, pass it.
+ .when(client)
+ .read(any(String.class), any(Long.class), any(byte[].class),
+ any(Integer.class), any(Integer.class), any(String.class));
+
+ AbfsInputStream inputStream = getAbfsInputStream(client, "testReadAheadManagerForFailedReadAhead.txt");
+
+ queueReadAheads(inputStream);
+
+ // AbfsInputStream Read would have waited for the read-ahead for the requested offset
+ // as we are testing from ReadAheadManager directly, sleep for a sec to
+ // get the read ahead threads to complete
+ Thread.sleep(1000);
+
+ // if readAhead failed for specific offset, getBlock should
+ // throw exception from the ReadBuffer that failed within last 30 sec
+ intercept(IOException.class,
+ () -> ReadBufferManager.getBufferManager().getBlock(
+ inputStream,
+ 0,
+ 1 * KILOBYTE,
+ new byte[1 * KILOBYTE]));
+
+ // Only the 3 readAhead threads should have triggered client.read
+ verifyReadCallCount(client, 3);
+
+ // Stub returns success for the 4th read request, if ReadBuffers still
+ // persisted, ReadAheadManager getBlock would have returned exception.
+ checkEvictedStatus(inputStream, 0, false);
+ }
+
+ /**
+ * The test expects ReadAheadManager to return 0 receivedBytes when previous
+ * read ahead on the offset had failed and not throw exception received then.
+ * Also checks that the ReadBuffers are evicted as per the ReadBufferManager
+ * threshold criteria.
+ * @throws Exception
+ */
+ @Test
+ public void testReadAheadManagerForOlderReadAheadFailure() throws Exception {
+ AbfsClient client = getMockAbfsClient();
+ AbfsRestOperation successOp = getMockRestOp();
+
+ // Stub :
+ // First Read request leads to 3 readahead calls: Fail all 3 readahead-client.read()
+ // A second read request will see that readahead had failed for data in
+ // the requested offset range but also that its is an older readahead request.
+ // System issue could have resolved by now, so attempt a new read only for the requested range.
+ doThrow(new TimeoutException("Internal Server error for RAH-X"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-X"))
+ .doThrow(new TimeoutException("Internal Server error for RAH-X"))
+ .doReturn(successOp) // pass the read for second read request
+ .doReturn(successOp) // pass success for post eviction test
+ .when(client)
+ .read(any(String.class), any(Long.class), any(byte[].class),
+ any(Integer.class), any(Integer.class), any(String.class));
+
+ AbfsInputStream inputStream = getAbfsInputStream(client, "testReadAheadManagerForOlderReadAheadFailure.txt");
+
+ queueReadAheads(inputStream);
+
+ // AbfsInputStream Read would have waited for the read-ahead for the requested offset
+ // as we are testing from ReadAheadManager directly, sleep for 30 secs so that
+ // read buffer qualifies for to be an old buffer
+ Thread.sleep(ReadBufferManager.getBufferManager().getThresholdAgeMilliseconds());
+
+ // Only the 3 readAhead threads should have triggered client.read
+ verifyReadCallCount(client, 3);
+
+ // getBlock from a new read request should return 0 if there is a failure
+ // 30 sec before in read ahead buffer for respective offset.
+ int bytesRead = ReadBufferManager.getBufferManager().getBlock(
+ inputStream,
+ 1 * KILOBYTE,
+ 1 * KILOBYTE,
+ new byte[1 * KILOBYTE]);
+ Assert.assertTrue("bytesRead should be zero when previously read "
Review comment:
assertEquals for all equality checks
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] goiri commented on a change in pull request #1898: Report
read-ahead error back
Posted by GitBox <gi...@apache.org>.
goiri commented on a change in pull request #1898: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r393791625
##########
File path: hadoop-tools/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/services/ReadBufferManager.java
##########
@@ -90,8 +94,10 @@ private ReadBufferManager() {
* @param stream The {@link AbfsInputStream} for which to do the read-ahead
* @param requestedOffset The offset in the file which shoukd be read
* @param requestedLength The length to read
+ * @param queueReadAheadRequestId unique queue request ID
*/
- void queueReadAhead(final AbfsInputStream stream, final long requestedOffset, final int requestedLength) {
+ void queueReadAhead(final AbfsInputStream stream, final long requestedOffset, final int requestedLength
Review comment:
80 char limit
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] hadoop-yetus commented on issue #1898: HADOOP-16852:
Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
hadoop-yetus commented on issue #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#issuecomment-609750216
:confetti_ball: **+1 overall**
| Vote | Subsystem | Runtime | Comment |
|:----:|----------:|--------:|:--------|
| +0 :ok: | reexec | 1m 27s | Docker mode activated. |
||| _ Prechecks _ |
| +1 :green_heart: | dupname | 0m 0s | No case conflicting files found. |
| +1 :green_heart: | @author | 0m 0s | The patch does not contain any @author tags. |
| +1 :green_heart: | test4tests | 0m 0s | The patch appears to include 1 new or modified test files. |
||| _ trunk Compile Tests _ |
| +1 :green_heart: | mvninstall | 21m 29s | trunk passed |
| +1 :green_heart: | compile | 0m 28s | trunk passed |
| +1 :green_heart: | checkstyle | 0m 21s | trunk passed |
| +1 :green_heart: | mvnsite | 0m 31s | trunk passed |
| +1 :green_heart: | shadedclient | 16m 30s | branch has no errors when building and testing our client artifacts. |
| +1 :green_heart: | javadoc | 0m 22s | trunk passed |
| +0 :ok: | spotbugs | 0m 49s | Used deprecated FindBugs config; considering switching to SpotBugs. |
| +1 :green_heart: | findbugs | 0m 48s | trunk passed |
| -0 :warning: | patch | 1m 5s | Used diff version of patch file. Binary files and potentially other changes not applied. Please rebase and squash commits if necessary. |
||| _ Patch Compile Tests _ |
| +1 :green_heart: | mvninstall | 0m 26s | the patch passed |
| +1 :green_heart: | compile | 0m 23s | the patch passed |
| +1 :green_heart: | javac | 0m 23s | the patch passed |
| -0 :warning: | checkstyle | 0m 14s | hadoop-tools/hadoop-azure: The patch generated 1 new + 0 unchanged - 0 fixed = 1 total (was 0) |
| +1 :green_heart: | mvnsite | 0m 25s | the patch passed |
| +1 :green_heart: | whitespace | 0m 0s | The patch has no whitespace issues. |
| +1 :green_heart: | shadedclient | 15m 26s | patch has no errors when building and testing our client artifacts. |
| +1 :green_heart: | javadoc | 0m 20s | the patch passed |
| +1 :green_heart: | findbugs | 0m 53s | the patch passed |
||| _ Other Tests _ |
| +1 :green_heart: | unit | 1m 13s | hadoop-azure in the patch passed. |
| +1 :green_heart: | asflicense | 0m 28s | The patch does not generate ASF License warnings. |
| | | 63m 16s | |
| Subsystem | Report/Notes |
|----------:|:-------------|
| Docker | Client=19.03.8 Server=19.03.8 base: https://builds.apache.org/job/hadoop-multibranch/job/PR-1898/7/artifact/out/Dockerfile |
| GITHUB PR | https://github.com/apache/hadoop/pull/1898 |
| JIRA Issue | HADOOP-16852 |
| Optional Tests | dupname asflicense compile javac javadoc mvninstall mvnsite unit shadedclient findbugs checkstyle |
| uname | Linux 46c06890e97a 4.15.0-74-generic #84-Ubuntu SMP Thu Dec 19 08:06:28 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux |
| Build tool | maven |
| Personality | personality/hadoop.sh |
| git revision | trunk / ab7495d |
| Default Java | 1.8.0_242 |
| checkstyle | https://builds.apache.org/job/hadoop-multibranch/job/PR-1898/7/artifact/out/diff-checkstyle-hadoop-tools_hadoop-azure.txt |
| Test Results | https://builds.apache.org/job/hadoop-multibranch/job/PR-1898/7/testReport/ |
| Max. process+thread count | 308 (vs. ulimit of 5500) |
| modules | C: hadoop-tools/hadoop-azure U: hadoop-tools/hadoop-azure |
| Console output | https://builds.apache.org/job/hadoop-multibranch/job/PR-1898/7/console |
| versions | git=2.7.4 maven=3.3.9 findbugs=3.1.0-RC1 |
| Powered by | Apache Yetus 0.11.1 https://yetus.apache.org |
This message was automatically generated.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] hadoop-yetus commented on issue #1898: HADOOP-16852:
Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
hadoop-yetus commented on issue #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#issuecomment-610182688
:confetti_ball: **+1 overall**
| Vote | Subsystem | Runtime | Comment |
|:----:|----------:|--------:|:--------|
| +0 :ok: | reexec | 1m 28s | Docker mode activated. |
||| _ Prechecks _ |
| +1 :green_heart: | dupname | 0m 0s | No case conflicting files found. |
| +1 :green_heart: | @author | 0m 0s | The patch does not contain any @author tags. |
| +1 :green_heart: | test4tests | 0m 0s | The patch appears to include 1 new or modified test files. |
||| _ trunk Compile Tests _ |
| +1 :green_heart: | mvninstall | 22m 15s | trunk passed |
| +1 :green_heart: | compile | 0m 28s | trunk passed |
| +1 :green_heart: | checkstyle | 0m 21s | trunk passed |
| +1 :green_heart: | mvnsite | 0m 29s | trunk passed |
| +1 :green_heart: | shadedclient | 16m 35s | branch has no errors when building and testing our client artifacts. |
| +1 :green_heart: | javadoc | 0m 22s | trunk passed |
| +0 :ok: | spotbugs | 0m 57s | Used deprecated FindBugs config; considering switching to SpotBugs. |
| +1 :green_heart: | findbugs | 0m 53s | trunk passed |
| -0 :warning: | patch | 1m 13s | Used diff version of patch file. Binary files and potentially other changes not applied. Please rebase and squash commits if necessary. |
||| _ Patch Compile Tests _ |
| +1 :green_heart: | mvninstall | 0m 27s | the patch passed |
| +1 :green_heart: | compile | 0m 22s | the patch passed |
| +1 :green_heart: | javac | 0m 22s | the patch passed |
| +1 :green_heart: | checkstyle | 0m 16s | the patch passed |
| +1 :green_heart: | mvnsite | 0m 25s | the patch passed |
| +1 :green_heart: | whitespace | 0m 0s | The patch has no whitespace issues. |
| +1 :green_heart: | shadedclient | 15m 36s | patch has no errors when building and testing our client artifacts. |
| +1 :green_heart: | javadoc | 0m 19s | the patch passed |
| +1 :green_heart: | findbugs | 0m 53s | the patch passed |
||| _ Other Tests _ |
| +1 :green_heart: | unit | 1m 24s | hadoop-azure in the patch passed. |
| +1 :green_heart: | asflicense | 0m 28s | The patch does not generate ASF License warnings. |
| | | 64m 38s | |
| Subsystem | Report/Notes |
|----------:|:-------------|
| Docker | Client=19.03.8 Server=19.03.8 base: https://builds.apache.org/job/hadoop-multibranch/job/PR-1898/8/artifact/out/Dockerfile |
| GITHUB PR | https://github.com/apache/hadoop/pull/1898 |
| JIRA Issue | HADOOP-16852 |
| Optional Tests | dupname asflicense compile javac javadoc mvninstall mvnsite unit shadedclient findbugs checkstyle |
| uname | Linux 62d9b5860a70 4.15.0-74-generic #84-Ubuntu SMP Thu Dec 19 08:06:28 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux |
| Build tool | maven |
| Personality | personality/hadoop.sh |
| git revision | trunk / 0b855b9 |
| Default Java | 1.8.0_242 |
| Test Results | https://builds.apache.org/job/hadoop-multibranch/job/PR-1898/8/testReport/ |
| Max. process+thread count | 307 (vs. ulimit of 5500) |
| modules | C: hadoop-tools/hadoop-azure U: hadoop-tools/hadoop-azure |
| Console output | https://builds.apache.org/job/hadoop-multibranch/job/PR-1898/8/console |
| versions | git=2.7.4 maven=3.3.9 findbugs=3.1.0-RC1 |
| Powered by | Apache Yetus 0.11.1 https://yetus.apache.org |
This message was automatically generated.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] hadoop-yetus commented on issue #1898: HADOOP-16852:
Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
hadoop-yetus commented on issue #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#issuecomment-605975866
:confetti_ball: **+1 overall**
| Vote | Subsystem | Runtime | Comment |
|:----:|----------:|--------:|:--------|
| +0 :ok: | reexec | 1m 32s | Docker mode activated. |
||| _ Prechecks _ |
| +1 :green_heart: | dupname | 0m 0s | No case conflicting files found. |
| +1 :green_heart: | @author | 0m 0s | The patch does not contain any @author tags. |
| +1 :green_heart: | test4tests | 0m 0s | The patch appears to include 1 new or modified test files. |
||| _ trunk Compile Tests _ |
| +1 :green_heart: | mvninstall | 22m 1s | trunk passed |
| +1 :green_heart: | compile | 0m 27s | trunk passed |
| +1 :green_heart: | checkstyle | 0m 21s | trunk passed |
| +1 :green_heart: | mvnsite | 0m 30s | trunk passed |
| +1 :green_heart: | shadedclient | 16m 28s | branch has no errors when building and testing our client artifacts. |
| +1 :green_heart: | javadoc | 0m 23s | trunk passed |
| +0 :ok: | spotbugs | 0m 52s | Used deprecated FindBugs config; considering switching to SpotBugs. |
| +1 :green_heart: | findbugs | 0m 50s | trunk passed |
| -0 :warning: | patch | 1m 9s | Used diff version of patch file. Binary files and potentially other changes not applied. Please rebase and squash commits if necessary. |
||| _ Patch Compile Tests _ |
| +1 :green_heart: | mvninstall | 0m 28s | the patch passed |
| +1 :green_heart: | compile | 0m 23s | the patch passed |
| +1 :green_heart: | javac | 0m 23s | the patch passed |
| -0 :warning: | checkstyle | 0m 16s | hadoop-tools/hadoop-azure: The patch generated 26 new + 0 unchanged - 0 fixed = 26 total (was 0) |
| +1 :green_heart: | mvnsite | 0m 25s | the patch passed |
| +1 :green_heart: | whitespace | 0m 0s | The patch has no whitespace issues. |
| +1 :green_heart: | shadedclient | 17m 21s | patch has no errors when building and testing our client artifacts. |
| +1 :green_heart: | javadoc | 0m 24s | the patch passed |
| +1 :green_heart: | findbugs | 0m 59s | the patch passed |
||| _ Other Tests _ |
| +1 :green_heart: | unit | 1m 32s | hadoop-azure in the patch passed. |
| +1 :green_heart: | asflicense | 0m 38s | The patch does not generate ASF License warnings. |
| | | 66m 40s | |
| Subsystem | Report/Notes |
|----------:|:-------------|
| Docker | Client=19.03.8 Server=19.03.8 base: https://builds.apache.org/job/hadoop-multibranch/job/PR-1898/5/artifact/out/Dockerfile |
| GITHUB PR | https://github.com/apache/hadoop/pull/1898 |
| JIRA Issue | HADOOP-16852 |
| Optional Tests | dupname asflicense compile javac javadoc mvninstall mvnsite unit shadedclient findbugs checkstyle |
| uname | Linux 015a48eb2c98 4.15.0-74-generic #84-Ubuntu SMP Thu Dec 19 08:06:28 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux |
| Build tool | maven |
| Personality | personality/hadoop.sh |
| git revision | trunk / 960c9eb |
| Default Java | 1.8.0_242 |
| checkstyle | https://builds.apache.org/job/hadoop-multibranch/job/PR-1898/5/artifact/out/diff-checkstyle-hadoop-tools_hadoop-azure.txt |
| Test Results | https://builds.apache.org/job/hadoop-multibranch/job/PR-1898/5/testReport/ |
| Max. process+thread count | 308 (vs. ulimit of 5500) |
| modules | C: hadoop-tools/hadoop-azure U: hadoop-tools/hadoop-azure |
| Console output | https://builds.apache.org/job/hadoop-multibranch/job/PR-1898/5/console |
| versions | git=2.7.4 maven=3.3.9 findbugs=3.1.0-RC1 |
| Powered by | Apache Yetus 0.11.1 https://yetus.apache.org |
This message was automatically generated.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] snvijaya commented on issue #1898: HADOOP-16852: Report
read-ahead error back
Posted by GitBox <gi...@apache.org>.
snvijaya commented on issue #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#issuecomment-607010216
@steveloughran - Can you please help review this PR ?
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] hadoop-yetus commented on issue #1898: HADOOP-16852:
Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
hadoop-yetus commented on issue #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#issuecomment-609696019
:broken_heart: **-1 overall**
| Vote | Subsystem | Runtime | Comment |
|:----:|----------:|--------:|:--------|
| +0 :ok: | reexec | 0m 0s | Docker mode activated. |
| -1 :x: | patch | 0m 5s | https://github.com/apache/hadoop/pull/1898 does not apply to trunk. Rebase required? Wrong Branch? See https://wiki.apache.org/hadoop/HowToContribute for help. |
| Subsystem | Report/Notes |
|----------:|:-------------|
| GITHUB PR | https://github.com/apache/hadoop/pull/1898 |
| JIRA Issue | HADOOP-16852 |
| Console output | https://builds.apache.org/job/hadoop-multibranch/job/PR-1898/6/console |
| versions | git=2.17.1 |
| Powered by | Apache Yetus 0.11.1 https://yetus.apache.org |
This message was automatically generated.
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org
[GitHub] [hadoop] goiri commented on a change in pull request #1898:
HADOOP-16852: Report read-ahead error back
Posted by GitBox <gi...@apache.org>.
goiri commented on a change in pull request #1898: HADOOP-16852: Report read-ahead error back
URL: https://github.com/apache/hadoop/pull/1898#discussion_r395128629
##########
File path: hadoop-tools/hadoop-azure/src/test/java/org/apache/hadoop/fs/azurebfs/services/TestAbfsInputStream.java
##########
@@ -0,0 +1,438 @@
+/**
+ * Licensed to the Apache Software Foundation (ASF) under one
+ * or more contributor license agreements. See the NOTICE file
+ * distributed with this work for additional information
+ * regarding copyright ownership. The ASF licenses this file
+ * to you under the Apache License, Version 2.0 (the
+ * "License"); you may not use this file except in compliance
+ * with the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.hadoop.fs.azurebfs.services;
+
+import java.io.IOException;
+import java.util.UUID;
+
+import org.junit.Assert;
+import org.junit.Test;
+
+import org.apache.hadoop.fs.azurebfs.AbstractAbfsIntegrationTest;
+import org.apache.hadoop.fs.azurebfs.contracts.exceptions.AzureBlobFileSystemException;
+import org.apache.hadoop.fs.azurebfs.contracts.exceptions.TimeoutException;
+
+import static java.util.UUID.randomUUID;
+import static org.mockito.ArgumentMatchers.any;
+import static org.mockito.Mockito.doReturn;
+import static org.mockito.Mockito.doThrow;
+import static org.mockito.Mockito.mock;
+import static org.mockito.Mockito.times;
+import static org.mockito.Mockito.verify;
+import static org.mockito.Mockito.when;
+import static org.apache.hadoop.test.LambdaTestUtils.intercept;
+import static org.apache.hadoop.fs.azurebfs.constants.AbfsHttpConstants.FORWARD_SLASH;
+
+/**
+ * Unit test AbfsInputStream.
+ */
+public class TestAbfsInputStream extends
Review comment:
I was referring to check using LogCapturer
----------------------------------------------------------------
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:
users@infra.apache.org
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org