Sunil G commented on YARN-4138:

HI [~mding]
In the same case, if we 2 incremental requests, like 2G --> 4G --> 8G. Is it 
possible to have a success case where the container size reached 4G at any 
intermediate time?. Because from RMs point of view, last incremental request is 
to reach 8GB. In NM, it may have reached 4G already. And here if we get a 
decrease request to 6G, by this logic RM container size will it fall down to 
2G. (But in real, 4G was attained in NM). Kindly let me know if I missed 
something in this analysis.

> 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
> 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