HADOOP-19448: [ABFS][FNSOverBlob][Optimizations] Reduce Network Calls…

… In Create and Mkdir Flow (#7353) Contributed by Anmol Asrani Signed off by: Anuj Modi<[email protected]>
apache · Feb 7, 2025 · 73ac0b9 · 73ac0b9
1 parent d552670
commit 73ac0b9
Show file tree

Hide file tree

Showing 15 changed files with 1,839 additions and 991 deletions.
diff --git a/...ls/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/AzureBlobFileSystemStore.java b/...ls/hadoop-azure/src/main/java/org/apache/hadoop/fs/azurebfs/AzureBlobFileSystemStore.java
@@ -70,7 +70,6 @@
 import org.apache.hadoop.fs.azurebfs.constants.HttpHeaderConfigurations;
 import org.apache.hadoop.fs.azurebfs.contracts.exceptions.AbfsRestOperationException;
 import org.apache.hadoop.fs.azurebfs.contracts.exceptions.AzureBlobFileSystemException;
-import org.apache.hadoop.fs.azurebfs.contracts.exceptions.ConcurrentWriteOperationDetectedException;
 import org.apache.hadoop.fs.azurebfs.contracts.exceptions.FileSystemOperationUnhandledException;
 import org.apache.hadoop.fs.azurebfs.contracts.exceptions.InvalidAbfsRestOperationException;
 import org.apache.hadoop.fs.azurebfs.contracts.exceptions.InvalidFileSystemPropertyException;
@@ -762,53 +761,8 @@ private AbfsRestOperation conditionalCreateOverwriteFile(final String relativePa
       final TracingContext tracingContext) throws IOException {
     AbfsRestOperation op;
     AbfsClient createClient = getClientHandler().getIngressClient();
-    try {
-      // Trigger a create with overwrite=false first so that eTag fetch can be
-      // avoided for cases when no pre-existing file is present (major portion
-      // of create file traffic falls into the case of no pre-existing file).
-      op = createClient.createPath(relativePath, true, false, permissions,
-          isAppendBlob, null, contextEncryptionAdapter, tracingContext);
-
-    } catch (AbfsRestOperationException e) {
-      if (e.getStatusCode() == HttpURLConnection.HTTP_CONFLICT) {
-        // File pre-exists, fetch eTag
-        try {
-          op = getClient().getPathStatus(relativePath, false, tracingContext, null);
-        } catch (AbfsRestOperationException ex) {
-          if (ex.getStatusCode() == HttpURLConnection.HTTP_NOT_FOUND) {
-            // Is a parallel access case, as file which was found to be
-            // present went missing by this request.
-            throw new ConcurrentWriteOperationDetectedException(
-                "Parallel access to the create path detected. Failing request "
-                    + "to honor single writer semantics");
-          } else {
-            throw ex;
-          }
-        }
-
-        String eTag = extractEtagHeader(op.getResult());
-
-        try {
-          // overwrite only if eTag matches with the file properties fetched befpre
-          op = createClient.createPath(relativePath, true, true, permissions,
-              isAppendBlob, eTag, contextEncryptionAdapter, tracingContext);
-        } catch (AbfsRestOperationException ex) {
-          if (ex.getStatusCode() == HttpURLConnection.HTTP_PRECON_FAILED) {
-            // Is a parallel access case, as file with eTag was just queried
-            // and precondition failure can happen only when another file with
-            // different etag got created.
-            throw new ConcurrentWriteOperationDetectedException(
-                "Parallel access to the create path detected. Failing request "
-                    + "to honor single writer semantics");
-          } else {
-            throw ex;
-          }
-        }
-      } else {
-        throw e;
-      }
-    }
-
+    op = createClient.conditionalCreateOverwriteFile(relativePath, statistics,
+        permissions, isAppendBlob, contextEncryptionAdapter, tracingContext);
     return op;
   }
 

diff --git a/...he/hadoop/fs/azurebfs/contracts/exceptions/ConcurrentWriteOperationDetectedException.java b/...he/hadoop/fs/azurebfs/contracts/exceptions/ConcurrentWriteOperationDetectedException.java
@@ -20,12 +20,29 @@
 
 /**
  * Thrown when a concurrent write operation is detected.
+ * This exception is used to indicate that parallel access to the create path
+ * has been detected, which violates the single writer semantics.
  */
 @org.apache.hadoop.classification.InterfaceAudience.Public
 @org.apache.hadoop.classification.InterfaceStability.Evolving
 public class ConcurrentWriteOperationDetectedException
     extends AzureBlobFileSystemException {
 
+  private static final String ERROR_MESSAGE = "Parallel access to the create path detected. Failing request "
+      + "to honor single writer semantics";
+
+  /**
+   * Constructs a new ConcurrentWriteOperationDetectedException with a default error message.
+   */
+  public ConcurrentWriteOperationDetectedException() {
+    super(ERROR_MESSAGE);
+  }
+
+  /**
+   * Constructs a new ConcurrentWriteOperationDetectedException with the specified error message.
+   *
+   * @param message the detail message.
+   */
   public ConcurrentWriteOperationDetectedException(String message) {
     super(message);
   }