[jira] [Created] (SPARK-13001) Coarse-grained Mesos scheduler should reject offers for longer period of time when reached max cores

["Sebastien Rainville (JIRA)" <ji...@apache.org>]

Sebastien Rainville created SPARK-13001:
-------------------------------------------

             Summary: Coarse-grained Mesos scheduler should reject offers for longer period of time when reached max cores
                 Key: SPARK-13001
                 URL: https://issues.apache.org/jira/browse/SPARK-13001
             Project: Spark
          Issue Type: Improvement
          Components: Mesos, Scheduler
    Affects Versions: 1.6.0, 1.5.2, 1.5.1, 1.5.0
            Reporter: Sebastien Rainville


Similar to https://issues.apache.org/jira/browse/SPARK-10471 but for "reached max cores".

The coarse-grained Mesos scheduler accepts every offer that match the requirements until it reaches "spark.cores.max", at which point it will reject every offer. Even though spark will never launch tasks on these offers, mesos keeps sending the same offers again every 5 seconds making them unavailable for other frameworks.

Spark should reject those offers for a longer period of time to prevent offer starvation when running a lot of frameworks.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org