[
https://issues.apache.org/jira/browse/FLINK-7294?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16163147#comment-16163147
]
ASF GitHub Bot commented on FLINK-7294:
---------------------------------------
Github user EronWright commented on the issue:
https://github.com/apache/flink/pull/4622
@bbayani thanks very much for this contribution. Please confirm my
understanding of the issue and verification. It appears that the resources
that are requested via `TaskInfo` must reference the framework role, otherwise
the resources are actually drawn from the unreserved pool. Your cluster is
configured such that all resources are reserved for a certain role, and so
Flink is unable to launch without this fix despite being configured to use the
role.
Have you verified correct behavior in the case where Flink's role is not
configured (and there is unreserved capacity)? Is the `TaskInfo` role set to
`*` and is that correct?
> mesos.resourcemanager.framework.role not working
> ------------------------------------------------
>
> Key: FLINK-7294
> URL: https://issues.apache.org/jira/browse/FLINK-7294
> Project: Flink
> Issue Type: Bug
> Components: Mesos
> Affects Versions: 1.3.1
> Reporter: Bhumika Bayani
> Assignee: Eron Wright
> Priority: Critical
>
> I am using the above said setting in flink-conf.yaml
> e.g.
> mesos.resourcemanager.framework.role: mesos_role_tasks
> I see a flink-scheduler registered in mesos/frameworks tab with above said
> role.
> But the scheduler fails to launch any tasks inspite of getting
> resource-offers from mesos-agents with correct role.
> The error seen is:
> {code}
> 2017-07-28 13:23:00,683 INFO
> org.apache.flink.mesos.runtime.clusterframework.MesosFlinkResourceManager -
> Mesos task taskmanager-03768 failed, with a TaskManager in launch or
> registration. State: TASK_ERROR Reason: REASON_TASK_INVALID (Task uses more
> resources cpus(\*):1; mem(\*):1024; ports(\*):[4006-4007] than available
> cpus(mesos_role_tasks):7.4; mem(mesos_role_tasks):45876;
> ports(mesos_role_tasks):[4002-4129, 4131-4380, 4382-4809, 4811-4957,
> 4959-4966, 4968-4979, 4981-5049, 31000-31196, 31198-31431, 31433-31607,
> 31609-32000]; ephemeral_storage(mesos_role_tasks):37662;
> efs_storage(mesos_role_tasks):8.79609e+12; disk(mesos_role_tasks):5115)
> {code}
> The request is made for resources with * role. We do not have mesos running
> anywhere with * role. Thus task manager never come up.
> Am I missing any configuration?
> I am using flink version 1.3.1
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
