[ 
https://issues.apache.org/jira/browse/FLINK-10148?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16632578#comment-16632578
 ] 

David Haglund commented on FLINK-10148:
---------------------------------------

This issue seems to be fixed in 1.6.1 and with web.timeout set to something to 
longer than 10 s. Setting web.timeout in version 1.6.0 was not sufficient. 

> flink cancel withSavepoint does not seem to use akka properties from 
> flink-conf.yaml with Flink 1.6.0 as with flink 1.4.x
> -------------------------------------------------------------------------------------------------------------------------
>
>                 Key: FLINK-10148
>                 URL: https://issues.apache.org/jira/browse/FLINK-10148
>             Project: Flink
>          Issue Type: Bug
>          Components: Client
>    Affects Versions: 1.6.0
>         Environment: Flink 1.6.0
> Docker image based on:
>  flink:1.6.0-hadoop27-scala_2.11
>   
> The docker instances run in Kubernetes. 
> One pod for the jobmanager, One pod for the taskmanager and another pod to 
> submit/cancel jobs. 
>            Reporter: David Haglund
>            Priority: Minor
>         Attachments: flink-1.6.0-cancel-timeout.txt, flink-conf.yaml
>
>
> I get an error when I'm trying to cancel a job a flink job with a savepoint. 
> The problem is that it takes longer than 10 seconds to take the save point so 
> I get an exception "Ask timed out on [Actor[akka://...]] after [10000 ms].", 
> I have attached the full error message in the file 
> link-1.6.0-cancel-timeout.txt.
> I have set akka.client.timeout and akka.ask.timeout in flink-conf.yaml but 
> they do not seem to be used. 
> It worked fine to set  akka.client.timeout and akka.ask.timeout in 
> flink-conf.yaml  with flink 1.4.2 but it does not seem to work with flink 
> 1.6.0.
>  
>  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to