Ngone51 commented on a change in pull request #33451:
URL: https://github.com/apache/spark/pull/33451#discussion_r676334962
##########
File path:
common/network-shuffle/src/main/java/org/apache/spark/network/shuffle/ExternalShuffleBlockResolver.java
##########
@@ -374,6 +376,32 @@ public int removeBlocks(String appId, String execId,
String[] blockIds) {
.collect(Collectors.toMap(Pair::getKey, Pair::getValue));
}
+ /**
+ * Diagnose the possible cause of the shuffle data corruption by verify the
shuffle checksums
+ */
+ public Cause diagnoseShuffleBlockCorruption(
+ String appId,
+ String execId,
+ int shuffleId,
+ long mapId,
+ int reduceId,
+ long checksumByReader) {
+ ExecutorShuffleInfo executor = executors.get(new AppExecId(appId, execId));
+ String fileName = "shuffle_" + shuffleId + "_" + mapId + "_0.checksum";
+ File probeFile = ExecutorDiskUtils.getFile(
+ executor.localDirs,
+ executor.subDirsPerLocalDir,
+ fileName);
+ File parentFile = probeFile.getParentFile();
+
+ File[] checksumFiles = parentFile.listFiles(f ->
f.getName().startsWith(fileName));
+ assert checksumFiles.length == 1;
Review comment:
> Is it because it was decided to append the checksum algorithm to the
file name and the exact name of the file is unknown?
Exactly. I have switched to use `Files.newDirectoryStream` instead.
--
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]
