[jira] [Commented] (SPARK-41497) Accumulator undercounting in the case of retry task with rdd cache

[ https://issues.apache.org/jira/browse/SPARK-41497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17696526#comment-17696526 ] Apache Spark commented on SPARK-41497: -- User 'ivoson' has created a pull request for this issue:

[jira] [Commented] (SPARK-41497) Accumulator undercounting in the case of retry task with rdd cache

[ https://issues.apache.org/jira/browse/SPARK-41497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17655807#comment-17655807 ] Apache Spark commented on SPARK-41497: -- User 'ivoson' has created a pull request for this issue:

[jira] [Commented] (SPARK-41497) Accumulator undercounting in the case of retry task with rdd cache

[ https://issues.apache.org/jira/browse/SPARK-41497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17655806#comment-17655806 ] Apache Spark commented on SPARK-41497: -- User 'ivoson' has created a pull request for this issue:

[jira] [Commented] (SPARK-41497) Accumulator undercounting in the case of retry task with rdd cache

[ https://issues.apache.org/jira/browse/SPARK-41497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17654590#comment-17654590 ] Mridul Muralidharan commented on SPARK-41497: - Sounds good [~Ngone51], thanks ! >

[jira] [Commented] (SPARK-41497) Accumulator undercounting in the case of retry task with rdd cache

[ https://issues.apache.org/jira/browse/SPARK-41497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17653781#comment-17653781 ] wuyi commented on SPARK-41497: -- > If I am not wrong, SQL makes very heavy use of accumulators, and so most

[jira] [Commented] (SPARK-41497) Accumulator undercounting in the case of retry task with rdd cache

[ https://issues.apache.org/jira/browse/SPARK-41497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17653648#comment-17653648 ] Juliusz Sompolski commented on SPARK-41497: --- Note that this issue leads to a correctness issue

[jira] [Commented] (SPARK-41497) Accumulator undercounting in the case of retry task with rdd cache

[ https://issues.apache.org/jira/browse/SPARK-41497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17646972#comment-17646972 ] Mridul Muralidharan commented on SPARK-41497: - [~Ngone51] Agree, that is what I was not sure

[jira] [Commented] (SPARK-41497) Accumulator undercounting in the case of retry task with rdd cache

[ https://issues.apache.org/jira/browse/SPARK-41497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17646971#comment-17646971 ] wuyi commented on SPARK-41497: -- I'm thinking if we could improve the improved Option 4 by changing the rdd

[jira] [Commented] (SPARK-41497) Accumulator undercounting in the case of retry task with rdd cache

[ https://issues.apache.org/jira/browse/SPARK-41497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17646969#comment-17646969 ] wuyi commented on SPARK-41497: -- > do we have a way to do that ?   [~mridulm80]  Currently, we only have

[jira] [Commented] (SPARK-41497) Accumulator undercounting in the case of retry task with rdd cache

[ https://issues.apache.org/jira/browse/SPARK-41497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17646962#comment-17646962 ] Mridul Muralidharan commented on SPARK-41497: - Agree, if we can determine that - do we have

[jira] [Commented] (SPARK-41497) Accumulator undercounting in the case of retry task with rdd cache

[ https://issues.apache.org/jira/browse/SPARK-41497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17646919#comment-17646919 ] wuyi commented on SPARK-41497: -- [~mridulm80]  For b) and c), shouldn't we allow T2 to use the result of

[jira] [Commented] (SPARK-41497) Accumulator undercounting in the case of retry task with rdd cache

[ https://issues.apache.org/jira/browse/SPARK-41497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17646825#comment-17646825 ] Mridul Muralidharan commented on SPARK-41497: - > For example, a task is constructed by

[jira] [Commented] (SPARK-41497) Accumulator undercounting in the case of retry task with rdd cache

[ https://issues.apache.org/jira/browse/SPARK-41497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17646658#comment-17646658 ] huangtengfei commented on SPARK-41497: -- I also think that option3/4(include the improved proposal

[jira] [Commented] (SPARK-41497) Accumulator undercounting in the case of retry task with rdd cache

[ https://issues.apache.org/jira/browse/SPARK-41497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17646639#comment-17646639 ] wuyi commented on SPARK-41497: -- [~mridulm80] Sounds like a better idea than option 4. But I think it still

[jira] [Commented] (SPARK-41497) Accumulator undercounting in the case of retry task with rdd cache

[ https://issues.apache.org/jira/browse/SPARK-41497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17646477#comment-17646477 ] Mridul Muralidharan commented on SPARK-41497: - Agree - there appears to be a bunch of

[jira] [Commented] (SPARK-41497) Accumulator undercounting in the case of retry task with rdd cache

[ https://issues.apache.org/jira/browse/SPARK-41497?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17646097#comment-17646097 ] wuyi commented on SPARK-41497: -- [~mridulm80] [~tgraves] [~attilapiros] [~ivoson] any good ideas? >