Github user zhaoyunjiong commented on a diff in the pull request: https://github.com/apache/spark/pull/14887#discussion_r77383971 --- Diff: common/network-yarn/src/main/java/org/apache/spark/network/yarn/YarnShuffleService.java --- @@ -25,6 +25,8 @@ import com.google.common.collect.Lists; import org.apache.hadoop.conf.Configuration; import org.apache.hadoop.fs.Path; +import org.apache.hadoop.util.DiskChecker; --- End diff -- In our case, we have about 200 nodes which have the first disk broken, and spark application is not minority. We can't accept lost so much nodes just because the first disk was broken, also we can't accept let spark application fail randomly (I already patched 1.6 use same logic in our cluster so we can keep enable spark dynamic allocation). If no one see any value of the improvement, that's fine, I'll close this PR. Thank you for your time.
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org