Github user henryr commented on a diff in the pull request:

    https://github.com/apache/spark/pull/21068#discussion_r181912300
  
    --- Diff: 
resource-managers/yarn/src/main/scala/org/apache/spark/deploy/yarn/FailureWithinTimeIntervalTracker.scala
 ---
    @@ -0,0 +1,80 @@
    +/*
    + * 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.yarn
    +
    +import scala.collection.mutable
    +
    +import org.apache.spark.SparkConf
    +import org.apache.spark.internal.Logging
    +import org.apache.spark.util.{Clock, SystemClock}
    +
    +private[spark] class FailureWithinTimeIntervalTracker(sparkConf: 
SparkConf) extends Logging {
    +
    +  private var clock: Clock = new SystemClock
    +
    +  private val executorFailuresValidityInterval =
    +    
sparkConf.get(config.EXECUTOR_ATTEMPT_FAILURE_VALIDITY_INTERVAL_MS).getOrElse(-1L)
    +
    +  // Queue to store the timestamp of failed executors for each host
    +  private val failedExecutorsTimeStampsPerHost = mutable.Map[String, 
mutable.Queue[Long]]()
    +
    +  private val failedExecutorsTimeStamps = new mutable.Queue[Long]()
    +
    +  private def getNumFailuresWithinValidityInterval(
    --- End diff --
    
    It's not really clear what a 'validity interval' is. I think it means that 
only failures that have happened recently are considered valid? I think it 
would be clearer to call this `getNumFailuresSince()`, or 
`getRecentFailureCount()` or similar, and explicitly pass in the timestamp the 
caller wants to consider failures since. 
    
    If you do the latter, and drop the `endTime` argument, then you partly 
address the issue I raise below about how this mutates state, because 
`getRecentFailureCount()` suggests more clearly that it's expecting to take 
into account the current time. 


---

---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org

Reply via email to