[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17396691#comment-17396691
]
Till Rohrmann commented on FLINK-16069:
---
These are great results. Thanks a lot for improving the
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17395728#comment-17395728
]
Zhu Zhu commented on FLINK-16069:
-
Thanks for making the improvements and sharing the results!
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17393869#comment-17393869
]
Zhilong Hong commented on FLINK-16069:
--
After the optimization we've done in FLINK-23005, the
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17354757#comment-17354757
]
Zhu Zhu commented on FLINK-16069:
-
Even if the main thread can have the highest priority, GC problem can
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17354477#comment-17354477
]
Chesnay Schepler commented on FLINK-16069:
--
Have we ever experimented with thread priorities
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17354434#comment-17354434
]
Till Rohrmann commented on FLINK-16069:
---
Maybe one could start by taking {{#cpus/2}} as a default
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17354420#comment-17354420
]
Zhu Zhu commented on FLINK-16069:
-
Yes a dedicated {{serializationExecutor}} is an alternative. One
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17354373#comment-17354373
]
Till Rohrmann commented on FLINK-16069:
---
This is a good finding [~zhuzh]. I like the ideas.
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17352943#comment-17352943
]
Zhu Zhu commented on FLINK-16069:
-
Thanks for the suggestion and sorry for the late reply! [~trohrmann]
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17312220#comment-17312220
]
Till Rohrmann commented on FLINK-16069:
---
Maybe we should remove {{AkkaRpcService.getExecutor}} and
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17312218#comment-17312218
]
Till Rohrmann commented on FLINK-16069:
---
Could it be that we use somehow the {{ActorSystems}}
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17311960#comment-17311960
]
Zhu Zhu commented on FLINK-16069:
-
>From what I can see, heartbeat timeout happens because the scheduled
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17311563#comment-17311563
]
Till Rohrmann commented on FLINK-16069:
---
Thanks for providing the feedback [~zhuzh]. The result
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17291607#comment-17291607
]
Zhu Zhu commented on FLINK-16069:
-
I had a PoC to verify the idea to cache and reuse ShuffleDescriptor
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17291544#comment-17291544
]
Till Rohrmann commented on FLINK-16069:
---
This ticket has been touched by the backlog grooming
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17174053#comment-17174053
]
Yumeng Zhang commented on FLINK-16069:
--
[~zhuzh] Yes, we encountered the same problem when
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17173885#comment-17173885
]
Zhu Zhu commented on FLINK-16069:
-
Hi Yumeng, I think we have not reached a consensus. I once did a PoC
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17173036#comment-17173036
]
Yumeng Zhang commented on FLINK-16069:
--
Hello guys, do you reach a consensus about how to optimize
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17045531#comment-17045531
]
Zhu Zhu commented on FLINK-16069:
-
>> Is there anything in the ShuffleDescriptor that would be different
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17045395#comment-17045395
]
Stephan Ewen commented on FLINK-16069:
--
"each InputGateDeploymentDescriptor of B instances should
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17045135#comment-17045135
]
Zhu Zhu commented on FLINK-16069:
-
Making building deployment descriptor more static seems not to be an
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17045107#comment-17045107
]
Yingjie Cao commented on FLINK-16069:
-
Maybe we can split the tasks to be deployed into multi mini
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17041812#comment-17041812
]
Stephan Ewen commented on FLINK-16069:
--
It it possible to make more of the deployment descriptor
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17039916#comment-17039916
]
Till Rohrmann commented on FLINK-16069:
---
Whenever we access state which is not immutable belonging
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17039191#comment-17039191
]
huweihua commented on FLINK-16069:
--
Hi, [~trohrmann], thanks for your time. Yes, The iteration over the
[
https://issues.apache.org/jira/browse/FLINK-16069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17038976#comment-17038976
]
Till Rohrmann commented on FLINK-16069:
---
Hi [~huwh], do you know what exactly is taking so long.
26 matches
Mail list logo