[jira] [Commented] (SPARK-31555) Improve cache block migration
[ https://issues.apache.org/jira/browse/SPARK-31555?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17243839#comment-17243839 ] Hyukjin Kwon commented on SPARK-31555: -- [~holden] and [~dongjoon], I switched it to 3.2.0 but feel free to edit if I got wrong. > Improve cache block migration > - > > Key: SPARK-31555 > URL: https://issues.apache.org/jira/browse/SPARK-31555 > Project: Spark > Issue Type: Sub-task > Components: Spark Core >Affects Versions: 3.1.0 >Reporter: Holden Karau >Priority: Major > > We should explore the following improvements to cache block migration: > 1) Peer selection (right now may overbalance on certain peers) > 2) Do we need to configure the number of blocks to be migrated at the same > time > 3) Are there any blocks we don't need to replicate (e.g. they are already > stored on the desired number of executors even once we remove the executors > slated for decommissioning). > 4) Do we want to prioritize migrating blocks with no replicas > 5) Log the attempt number for debugging > 6) Clarify the logic for determining the number of replicas > 7) Consider using TestUtils.waitUntilExecutorsUp in tests rather than count > to wait for the executors to come up. imho this is the least important. -- This message was sent by Atlassian Jira (v8.3.4#803005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-31555) Improve cache block migration
[ https://issues.apache.org/jira/browse/SPARK-31555?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17234923#comment-17234923 ] Dongjoon Hyun commented on SPARK-31555: --- Gentle ping because the target version is `3.1.0` in this JIRA. > Improve cache block migration > - > > Key: SPARK-31555 > URL: https://issues.apache.org/jira/browse/SPARK-31555 > Project: Spark > Issue Type: Sub-task > Components: Spark Core >Affects Versions: 3.1.0 >Reporter: Holden Karau >Priority: Major > > We should explore the following improvements to cache block migration: > 1) Peer selection (right now may overbalance on certain peers) > 2) Do we need to configure the number of blocks to be migrated at the same > time > 3) Are there any blocks we don't need to replicate (e.g. they are already > stored on the desired number of executors even once we remove the executors > slated for decommissioning). > 4) Do we want to prioritize migrating blocks with no replicas > 5) Log the attempt number for debugging > 6) Clarify the logic for determining the number of replicas > 7) Consider using TestUtils.waitUntilExecutorsUp in tests rather than count > to wait for the executors to come up. imho this is the least important. -- This message was sent by Atlassian Jira (v8.3.4#803005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-31555) Improve cache block migration
[ https://issues.apache.org/jira/browse/SPARK-31555?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17113736#comment-17113736 ] Holden Karau commented on SPARK-31555: -- Rocking. We're running into an issue in master with the tests so I'm going to take #7 as a seperate issue (the testutils) but otherwise have at it and let us know if you get stuck. > Improve cache block migration > - > > Key: SPARK-31555 > URL: https://issues.apache.org/jira/browse/SPARK-31555 > Project: Spark > Issue Type: Sub-task > Components: Spark Core >Affects Versions: 3.1.0 >Reporter: Holden Karau >Priority: Major > > We should explore the following improvements to cache block migration: > 1) Peer selection (right now may overbalance on certain peers) > 2) Do we need to configure the number of blocks to be migrated at the same > time > 3) Are there any blocks we don't need to replicate (e.g. they are already > stored on the desired number of executors even once we remove the executors > slated for decommissioning). > 4) Do we want to prioritize migrating blocks with no replicas > 5) Log the attempt number for debugging > 6) Clarify the logic for determining the number of replicas > 7) Consider using TestUtils.waitUntilExecutorsUp in tests rather than count > to wait for the executors to come up. imho this is the least important. -- This message was sent by Atlassian Jira (v8.3.4#803005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
[jira] [Commented] (SPARK-31555) Improve cache block migration
[ https://issues.apache.org/jira/browse/SPARK-31555?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17110732#comment-17110732 ] Dale Richardson commented on SPARK-31555: - Hi [~holden], happy to have a go at this. > Improve cache block migration > - > > Key: SPARK-31555 > URL: https://issues.apache.org/jira/browse/SPARK-31555 > Project: Spark > Issue Type: Sub-task > Components: Spark Core >Affects Versions: 3.1.0 >Reporter: Holden Karau >Priority: Major > > We should explore the following improvements to cache block migration: > 1) Peer selection (right now may overbalance on certain peers) > 2) Do we need to configure the number of blocks to be migrated at the same > time > 3) Are there any blocks we don't need to replicate (e.g. they are already > stored on the desired number of executors even once we remove the executors > slated for decommissioning). > 4) Do we want to prioritize migrating blocks with no replicas > 5) Log the attempt number for debugging > 6) Clarify the logic for determining the number of replicas > 7) Consider using TestUtils.waitUntilExecutorsUp in tests rather than count > to wait for the executors to come up. imho this is the least important. -- This message was sent by Atlassian Jira (v8.3.4#803005) - To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org
