[jira] [Commented] (SPARK-21682) Caching 100k-task RDD GC-kills driver (due to updatedBlockStatuses?)

2017-08-09 Thread DjvuLee (JIRA)
[ https://issues.apache.org/jira/browse/SPARK-21682?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16120959#comment-16120959 ] DjvuLee commented on SPARK-21682: - Yes, our company also faced with this scalability problem, the driver

[jira] [Commented] (SPARK-21682) Caching 100k-task RDD GC-kills driver (due to updatedBlockStatuses?)

2017-08-09 Thread Ryan Williams (JIRA)
[ https://issues.apache.org/jira/browse/SPARK-21682?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16120537#comment-16120537 ] Ryan Williams commented on SPARK-21682: --- bq. But do you really need to create so many partitions?

[jira] [Commented] (SPARK-21682) Caching 100k-task RDD GC-kills driver (due to updatedBlockStatuses?)

2017-08-09 Thread Shixiong Zhu (JIRA)
[ https://issues.apache.org/jira/browse/SPARK-21682?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16120415#comment-16120415 ] Shixiong Zhu commented on SPARK-21682: -- I agree that driver is a bottleneck. I already saw several

[jira] [Commented] (SPARK-21682) Caching 100k-task RDD GC-kills driver due to updatedBlockStatuses

2017-08-09 Thread Ryan Williams (JIRA)
[ https://issues.apache.org/jira/browse/SPARK-21682?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16120409#comment-16120409 ] Ryan Williams commented on SPARK-21682: --- Interestingly, I thought the {{updatedBlockStatuses}}