[ https://issues.apache.org/jira/browse/SPARK-48580?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
gaoyajun02 updated SPARK-48580: ------------------------------- Parent: SPARK-33235 Issue Type: Sub-task (was: Bug) > Add consistency check and fallback for mapIds in push-merged block meta > ----------------------------------------------------------------------- > > Key: SPARK-48580 > URL: https://issues.apache.org/jira/browse/SPARK-48580 > Project: Spark > Issue Type: Sub-task > Components: Shuffle > Affects Versions: 3.2.0, 3.3.0, 3.4.0, 3.5.0 > Reporter: gaoyajun02 > Priority: Major > Attachments: image-2024-06-11-10-19-57-227.png > > > When push-based shuffle enabled, 0.03% of the spark application in our > cluster experienced shuffle data loss. The metrics of Exchange as follows: > !image-2024-06-11-10-19-57-227.png|width=405,height=170! > We eventually found some WARN logs on the shuffle server: > > {code:java} > WARN shuffle-server-8-216 > org.apache.spark.network.shuffle.RemoteBlockPushResolver: Application > application_xxxx shuffleId 0 shuffleMergeId 0 reduceId 133 update to > index/meta failed{code} > > And analyzed the cause from the code: > The merge metadata obtained by the reduce side from the driver comes from the > {{mapTracker}} in the server's memory, while the actual reading of chunk data > is based on the records in the shuffle server's {{{}metaFile{}}}. There is no > consistency check between the two. -- This message was sent by Atlassian Jira (v8.20.10#820010) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org