[
https://issues.apache.org/jira/browse/YARN-1018?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13732244#comment-13732244
]
Sandy Ryza commented on YARN-1018:
----------------------------------
If I understand correctly, you're saying that someone might want to colocate
containers with other containers that have already been scheduled, so the first
can be anywhere, and the later ones need to be strictly on the same node(s) as
the first. While this is a situation that we should support, it seems to me
that it should require the special handling that you're suggesting. By the
semantics of relaxLocality, the first request is expected to behave
semantically differently than the ones that will come after it. It is not a
strictly local request, so setting the relaxLocality flag to false on it
doesn't make sense.
> prereq check for AMRMClient.ContainerRequest relaxLocality flag wrong
> ---------------------------------------------------------------------
>
> Key: YARN-1018
> URL: https://issues.apache.org/jira/browse/YARN-1018
> Project: Hadoop YARN
> Issue Type: Bug
> Components: client
> Affects Versions: 2.1.0-beta
> Reporter: Steve Loughran
> Priority: Minor
>
> Trying to create a container request with no racks/nodes and no relaxed
> priority fails
> {code}
> new AMRMClient.ContainerRequest(capability, null, null, 0, false);
> {code}
> expected: a container request.
> actual: stack trace saying I can't relax node locality.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
