Re: [PR] [SPARK-43301][CORE][SHUFFLE] BlockStoreClient getHostLocalDirs RPC supports IOException retry [spark]

["mridulm (via GitHub)" <gi...@apache.org>]

mridulm commented on code in PR #46805:
URL: https://github.com/apache/spark/pull/46805#discussion_r1743011753


##########
common/network-shuffle/src/main/java/org/apache/spark/network/shuffle/BlockStoreClient.java:
##########
@@ -161,6 +172,18 @@ public void getHostLocalDirs(
       String[] execIds,
       CompletableFuture<Map<String, String[]>> hostLocalDirsCompletable) {
     checkInit();
+    int maxRetries = transportConf.maxIORetries();
+    int retryWaitTime = transportConf.ioRetryWaitTimeMs();
+    boolean enableSaslRetries = transportConf.enableSaslRetries();
+    retry(0, 0, maxRetries, enableSaslRetries, retryWaitTime,
+      () -> getHostLocalDirsInternal(host, port, execIds), hostLocalDirsCompletable);

Review Comment:
   This can wait for an extended duration if there is an existing call blocking on `getHostLocalDirsInternal`.
   My [query here](https://github.com/apache/spark/pull/46805/files#r1643785281)  was not to remove the threadpool and replace with single threaded executor - but rather understand better how it should be sized.



-- 
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: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


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