sarutak commented on code in PR #57387:
URL: https://github.com/apache/spark/pull/57387#discussion_r3629896137


##########
core/src/main/scala/org/apache/spark/deploy/security/UserCredentialManager.scala:
##########
@@ -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.spark.deploy.security
+
+import java.io.{ByteArrayOutputStream, ObjectInputFilter, ObjectInputStream, 
ObjectOutputStream}
+import java.net.URI
+import java.nio.file.Paths
+import java.time.Instant
+import java.util.concurrent.{RejectedExecutionException, 
ScheduledExecutorService, TimeUnit}
+import java.util.concurrent.atomic.AtomicInteger
+
+import scala.collection.mutable
+import scala.jdk.CollectionConverters._
+
+import org.apache.spark.SparkConf
+import org.apache.spark.internal.Logging
+import org.apache.spark.internal.LogKeys
+import org.apache.spark.internal.config._
+import org.apache.spark.security._
+import org.apache.spark.ui.UIUtils
+import org.apache.spark.util.ThreadUtils
+
+/**
+ * Manager for OIDC-based credential propagation on the driver.
+ *
+ * This class is a sibling of [[HadoopDelegationTokenManager]] and handles the 
OIDC credential
+ * propagation path independently. Both managers run on independent threads 
and can both be
+ * active simultaneously (e.g., a cluster accessing HDFS via Kerberos and S3 
via OIDC).
+ *
+ * Responsibilities:
+ *   1. Reads the current identity token via a [[TokenIngestor]]
+ *   2. Calls [[CredentialProvider#resolve]] for each configured scheme
+ *   3. Serializes the resulting [[UserCredentials]] and invokes the 
propagation callback
+ *   4. Schedules renewal based on `min(identity token expiry, service 
credential expiry) -
+ *      safetyMargin`
+ *   5. Retries with exponential backoff on failure
+ *
+ * Started from `CoarseGrainedSchedulerBackend.start()` when
+ * `spark.security.credentials.enabled=true`, independently of
+ * `UserGroupInformation.isSecurityEnabled()`.
+ *
+ * Lifecycle: call `start()` exactly once, then `stop()` to shut down.
+ * Calling `start()` after `stop()` is not supported.
+ */
+private[spark] class UserCredentialManager(
+    sparkConf: SparkConf,
+    tokenIngestor: TokenIngestor,
+    onCredentialsUpdate: Array[Byte] => Unit) extends Logging {
+
+  private val safetyMargin = 
sparkConf.get(SECURITY_CREDENTIALS_RENEWAL_SAFETY_MARGIN)
+  private val minInterval = 
sparkConf.get(SECURITY_CREDENTIALS_RENEWAL_MIN_INTERVAL)
+
+  // Thread-safe counter for exponential backoff calculation.
+  // Accessed from both the caller of start() and the scheduled renewal thread.
+  private val consecutiveFailures = new AtomicInteger(0)
+  private val maxBackoffMs: Long = TimeUnit.MINUTES.toMillis(10)
+
+  @volatile private var renewalExecutor: ScheduledExecutorService = _
+
+  // Pre-compute the filtered credential configuration map (SparkConf is 
immutable).
+  private val credentialConfMap: java.util.Map[String, String] = 
sparkConf.getAll
+    .filter(_._1.startsWith("spark.security.credentials."))
+    .toMap.asJava
+
+  /**
+   * Start the credential manager. Acquires initial credentials and schedules 
renewal.
+   *
+   * The initial credential acquisition is fail-fast: if the token cannot be 
loaded or
+   * no credentials can be resolved, an exception is thrown. Subsequent 
renewal failures
+   * are handled with exponential backoff.
+   *
+   * @return The serialized initial [[UserCredentials]].
+   * @throws IllegalStateException if the initial credential acquisition fails.
+   */
+  def start(): Array[Byte] = {
+    require(renewalExecutor == null, "start() must not be called more than 
once")
+    renewalExecutor =
+      
ThreadUtils.newDaemonSingleThreadScheduledExecutor("user-credential-renewal")
+
+    // Initial acquisition is fail-fast (no retry/backoff).
+    // This ensures the application fails to start if credentials cannot be 
obtained,
+    // rather than running with null credentials.
+    val userContext = tokenIngestor.load()
+    if (!userContext.isPresent) {
+      throw new IllegalStateException(
+        "Failed to start UserCredentialManager: identity token file is missing 
or malformed. " +
+          s"Check ${SECURITY_CREDENTIALS_IDENTITY_TOKEN_FILE.key} 
configuration.")
+    }
+
+    val ctx = userContext.get()
+    logInfo(log"Initial identity token loaded for principal " +
+      log"${MDC(LogKeys.PRINCIPAL, ctx.getPrincipal)} " +
+      log"(issuer: ${MDC(LogKeys.URI, ctx.getIssuer)})")
+
+    val (credentials, earliestExpiry) = resolveCredentials(ctx)
+    val serialized = serializeUserCredentials(credentials)
+
+    // Propagate initial credentials
+    onCredentialsUpdate(serialized)
+
+    // Schedule first renewal
+    val renewalDelay = computeRenewalDelay(ctx, earliestExpiry)
+    scheduleRenewal(renewalDelay)
+
+    logInfo(log"Credential acquisition successful. Next renewal in " +
+      log"${MDC(LogKeys.TIME_UNITS, UIUtils.formatDuration(renewalDelay))}.")
+    serialized
+  }
+
+  def stop(): Unit = {
+    if (renewalExecutor != null) {
+      renewalExecutor.shutdownNow()
+    }
+  }
+
+  /**
+   * Scheduled renewal task: load identity token, resolve credentials, 
propagate,
+   * and schedule the next renewal. Failures are retried with exponential 
backoff.
+   */
+  private def renewCredentialsTask(): Unit = {
+    try {
+      val userContext = tokenIngestor.load()
+      if (!userContext.isPresent) {
+        throw new IllegalStateException(
+          "TokenIngestor returned empty - identity token file may be missing 
or malformed")
+      }
+
+      val ctx = userContext.get()
+      logInfo(log"Loaded identity token for principal " +
+        log"${MDC(LogKeys.PRINCIPAL, ctx.getPrincipal)} " +
+        log"(issuer: ${MDC(LogKeys.URI, ctx.getIssuer)})")
+
+      val (credentials, earliestExpiry) = resolveCredentials(ctx)
+      val serialized = serializeUserCredentials(credentials)
+
+      // Propagate credentials to executors. Errors here are logged separately
+      // so that credential-fetch success is not conflated with distribution 
failure.
+      try {
+        onCredentialsUpdate(serialized)
+      } catch {
+        case e: Exception =>
+          logWarning(log"Credentials were resolved successfully but failed to 
propagate " +
+            log"to executors. Executors will receive updated credentials on 
next renewal.", e)
+      }
+
+      // Reset backoff on successful credential resolution. This is 
intentionally done
+      // even if onCredentialsUpdate failed above: the backoff counter tracks 
credential
+      // *resolution* failures (STS/provider issues), not propagation failures.
+      // Propagation failures are transient and will be retried on next 
scheduled renewal.
+      consecutiveFailures.set(0)

Review Comment:
   For now, the design relies on Sub-task 5's `TaskDescription`-based 
credential delivery, which guarantees credentials are available before task 
execution regardless of RPC broadcast timing. This makes the broadcast a 
best-effort optimization for credential refresh during long-running tasks 
rather than a critical delivery path.
   
   If we observe propagation failures causing stale credentials in practice, a 
shorter propagation-only retry (skipping `tokenIngestor.load()` + 
`resolveCredentials()`) would be straightforward to add as a follow-up.



-- 
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.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to