sandflee commented on YARN-4138:

Hi, [~mding], consider such situation:

1)  AM request increase request to RM and received a token.
2)  AM sends increase request to NM and succed (this may take a long time 
because of AM logic or network problem or nm restart)
3)  RM token expires, and send decrease action to NM
4)  NM sends increase notify to RM, since token expires, ignored by RM

to AM, since there's no increase succ/expire by RM, resource are successfully 
increased, but to RM and NM, increase action are expired 

> Roll back container resource allocation after resource increase token expires
> -----------------------------------------------------------------------------
>                 Key: YARN-4138
>                 URL: https://issues.apache.org/jira/browse/YARN-4138
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: api, nodemanager, resourcemanager
>            Reporter: MENG DING
>            Assignee: MENG DING
>         Attachments: YARN-4138-YARN-1197.1.patch, YARN-4138-YARN-1197.2.patch
> In YARN-1651, after container resource increase token expires, the running 
> container is killed.
> This ticket will change the behavior such that when a container resource 
> increase token expires, the resource allocation of the container will be 
> reverted back to the value before the increase.

This message was sent by Atlassian JIRA

Reply via email to