Github user WeichenXu123 commented on the issue: https://github.com/apache/spark/pull/14333 @srowen I check the code about KMean `bcNewCenters` again, if we want to make sure the recovery of RDD will successful in any unexcepted case, we have to keep all the `bcNewCenters` generated in each loop, until the loop is done. because each loop the `costs:RDD` is build using preCosts:RDD, so that it became a RDD link. and the loop will and only will keep latest two RDDs being persisted. if the last two RDDs is broken, spark will need to rebuild them from the first RDD, in such case, each historical `bcNewCenters` generated will be used.
--- 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