deshanxiao opened a new pull request #25692: [SPARK-28987][CORE] DiskBlockManager#createTempShuffleBlock should skip read-only directory URL: https://github.com/apache/spark/pull/25692 ### What changes were proposed in this pull request? Skip read-only directory when pick the path of shuffle block file ### Why are the changes needed? DiskBlockManager#createTempShuffleBlock only considers the path which is not exist. I think we could check whether the path is writeable or not. It's resonable beacuse we invoke createTempShuffleBlock to create a new path to write files in it. It should be writeable. Exception Stack: ``` Caused by: org.apache.spark.SparkException: Job aborted due to stage failure: Task 1765 in stage 368592.0 failed 4 times, most recent failure: Lost task 1765.3 in stage 368592.0 (TID 66021932, test-hadoop-prc-st2808.bj, executor 251): java.io.FileNotFoundException: /home/work/hdd6/yarn/test-hadoop/nodemanager/usercache/sql_test/appcache/application_1560996968289_16320/blockmgr-14608b48-7efd-4fd3-b050-2ac9953390d4/1e/temp_shuffle_00c7b87f-d7ed-49f3-90e7-1c8358bcfd74 (No such file or directory) at java.io.FileOutputStream.open0(Native Method) at java.io.FileOutputStream.open(FileOutputStream.java:270) at java.io.FileOutputStream.<init>(FileOutputStream.java:213) at org.apache.spark.storage.DiskBlockObjectWriter.initialize(DiskBlockObjectWriter.scala:139) at org.apache.spark.storage.DiskBlockObjectWriter.open(DiskBlockObjectWriter.scala:150) at org.apache.spark.storage.DiskBlockObjectWriter.write(DiskBlockObjectWriter.scala:268) at org.apache.spark.shuffle.sort.BypassMergeSortShuffleWriter.write(BypassMergeSortShuffleWriter.java:159) at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:96) at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:53) at org.apache.spark.scheduler.Task.run(Task.scala:100) at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:335) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) at java.lang.Thread.run(Thread.java:748) Driver stacktrace: at org.apache.spark.scheduler.DAGScheduler.org$apache$spark$scheduler$DAGScheduler$$failJobAndIndependentStages(DAGScheduler.scala:1515) at org.apache.spark.scheduler.DAGScheduler$$anonfun$abortStage$1.apply(DAGScheduler.scala:1503) at org.apache.spark.scheduler.DAGScheduler$$anonfun$abortStage$1.apply(DAGScheduler.scala:1502) at scala.collection.mutable.ResizableArray$class.foreach(ResizableArray.scala:59) at scala.collection.mutable.ArrayBuffer.foreach(ArrayBuffer.scala:48) at org.apache.spark.scheduler.DAGScheduler.abortStage(DAGScheduler.scala:1502) at org.apache.spark.scheduler.DAGScheduler$$anonfun$handleTaskSetFailed$1.apply(DAGScheduler.scala:816) at org.apache.spark.scheduler.DAGScheduler$$anonfun$handleTaskSetFailed$1.apply(DAGScheduler.scala:816) at scala.Option.foreach(Option.scala:257) at org.apache.spark.scheduler.DAGScheduler.handleTaskSetFailed(DAGScheduler.scala:816) at org.apache.spark.scheduler.DAGSchedulerEventProcessLoop.doOnReceive(DAGScheduler.scala:1740) at org.apache.spark.scheduler.DAGSchedulerEventProcessLoop.onReceive(DAGScheduler.scala:1695) at org.apache.spark.scheduler.DAGSchedulerEventProcessLoop.onReceive(DAGScheduler.scala:1684) at org.apache.spark.util.EventLoop$$anon$1.run(EventLoop.scala:48) ``` ### Does this PR introduce any user-facing change? No ### How was this patch tested? Unit test
---------------------------------------------------------------- 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: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
