[
https://issues.apache.org/jira/browse/FLINK-21030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17278088#comment-17278088
]
Matthias commented on FLINK-21030:
--
{quote}1. Taking a savepoint
2. Stopping the job gracefully
[
https://issues.apache.org/jira/browse/FLINK-21030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17277880#comment-17277880
]
Till Rohrmann commented on FLINK-21030:
---
Yes, fully agreed. Restarting the
[
https://issues.apache.org/jira/browse/FLINK-21030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17277769#comment-17277769
]
Zhu Zhu commented on FLINK-21030:
-
I guess I had misunderstood the protocol of
[
https://issues.apache.org/jira/browse/FLINK-21030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17277736#comment-17277736
]
Matthias commented on FLINK-21030:
--
Thanks for the analysis, that makes sense. I went ahead with this
[
https://issues.apache.org/jira/browse/FLINK-21030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17277248#comment-17277248
]
Till Rohrmann commented on FLINK-21030:
---
I think the problem can and should be solved in the scope
[
https://issues.apache.org/jira/browse/FLINK-21030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17277054#comment-17277054
]
Till Rohrmann commented on FLINK-21030:
---
I am not entirely sure whether adjusting the checkpoint
[
https://issues.apache.org/jira/browse/FLINK-21030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17277024#comment-17277024
]
Matthias commented on FLINK-21030:
--
You're right: My proposal would affect the checkpoint creation as
[
https://issues.apache.org/jira/browse/FLINK-21030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17276965#comment-17276965
]
Yun Tang commented on FLINK-21030:
--
I think returning the complete future to checkpoint coordinator to
[
https://issues.apache.org/jira/browse/FLINK-21030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17276920#comment-17276920
]
Zhu Zhu commented on FLINK-21030:
-
Sounds good to me.
[~yunta] would you also take a look at [~mapohl]'s
[
https://issues.apache.org/jira/browse/FLINK-21030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17276909#comment-17276909
]
Matthias commented on FLINK-21030:
--
Yes, the notification of tasks happens right at the end of
[
https://issues.apache.org/jira/browse/FLINK-21030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17276846#comment-17276846
]
Zhu Zhu commented on FLINK-21030:
-
Thanks for the investigation! [~mapohl]
Does this mean that the
[
https://issues.apache.org/jira/browse/FLINK-21030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17276539#comment-17276539
]
Matthias commented on FLINK-21030:
--
I paired with [~chesnay] to look into the issue. We came up with a
[
https://issues.apache.org/jira/browse/FLINK-21030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17272186#comment-17272186
]
Till Rohrmann commented on FLINK-21030:
---
No objections from my side.
> Broken job restart for job
[
https://issues.apache.org/jira/browse/FLINK-21030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17272108#comment-17272108
]
Matthias commented on FLINK-21030:
--
Just to clarify: The expected behavior is then that the command
[
https://issues.apache.org/jira/browse/FLINK-21030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17268409#comment-17268409
]
Zhu Zhu commented on FLINK-21030:
-
Agreed to trigger a global failover to bring FINISHED tasks back to
[
https://issues.apache.org/jira/browse/FLINK-21030?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17267923#comment-17267923
]
Till Rohrmann commented on FLINK-21030:
---
Thanks for reporting this issue [~TheoD]. I agree that
16 matches
Mail list logo