[jira] [Commented] (FLINK-19596) Do not recover CompletedCheckpointStore on each failover

[ https://issues.apache.org/jira/browse/FLINK-19596?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17217724#comment-17217724 ] Roman Khachatryan commented on FLINK-19596: --- Thanks for the clarification [~wind_ljy] FYI,

[jira] [Commented] (FLINK-19596) Do not recover CompletedCheckpointStore on each failover

[ https://issues.apache.org/jira/browse/FLINK-19596?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17216590#comment-17216590 ] Jiayi Liao commented on FLINK-19596: [~roman_khachatryan] In the losing-regaining leadership case,

[jira] [Commented] (FLINK-19596) Do not recover CompletedCheckpointStore on each failover

[ https://issues.apache.org/jira/browse/FLINK-19596?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17216581#comment-17216581 ] Roman Khachatryan commented on FLINK-19596: --- Hi, I was thinking about the solution for a

[jira] [Commented] (FLINK-19596) Do not recover CompletedCheckpointStore on each failover

[ https://issues.apache.org/jira/browse/FLINK-19596?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17216563#comment-17216563 ] Till Rohrmann commented on FLINK-19596: --- Thanks for creating this issue [~wind_ljy]. I think you

[jira] [Commented] (FLINK-19596) Do not recover CompletedCheckpointStore on each failover

[ https://issues.apache.org/jira/browse/FLINK-19596?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17216541#comment-17216541 ] Jiayi Liao commented on FLINK-19596: [~yunta] I didn't know there's already a JIRA before. Thanks

[jira] [Commented] (FLINK-19596) Do not recover CompletedCheckpointStore on each failover

[ https://issues.apache.org/jira/browse/FLINK-19596?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17216530#comment-17216530 ] Yun Tang commented on FLINK-19596: -- [~wind_ljy] this issue duplicates with FLINK-6984 and I think this

[jira] [Commented] (FLINK-19596) Do not recover CompletedCheckpointStore on each failover

[ https://issues.apache.org/jira/browse/FLINK-19596?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17216520#comment-17216520 ] Zhijiang commented on FLINK-19596: -- I heard of this issue proposed long time ago from [~tangyun], maybe

[jira] [Commented] (FLINK-19596) Do not recover CompletedCheckpointStore on each failover

[ https://issues.apache.org/jira/browse/FLINK-19596?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17216517#comment-17216517 ] Zhijiang commented on FLINK-19596: -- Thanks for proposing this potential improvement [~wind_ljy]! +1

[jira] [Commented] (FLINK-19596) Do not recover CompletedCheckpointStore on each failover

[ https://issues.apache.org/jira/browse/FLINK-19596?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17215808#comment-17215808 ] Nicholas Jiang commented on FLINK-19596: [~wind_ljy], I agree with the point you mentioned.

[jira] [Commented] (FLINK-19596) Do not recover CompletedCheckpointStore on each failover

[ https://issues.apache.org/jira/browse/FLINK-19596?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17212851#comment-17212851 ] Jiayi Liao commented on FLINK-19596: Currently {{restoreLatestCheckpointedStateInternal}} would be