[
https://issues.apache.org/jira/browse/SPARK-2294?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14045433#comment-14045433
]
Mridul Muralidharan commented on SPARK-2294:
I agree; We should bump no locality pref and speculative tasks to NODE_LOCAL
level after NODE_LOCAL tasks have been scheduled (if available), and not check
for them at PROCESS_LOCAL max locality. So they get scheduled before RACK_LOCAL
but after NODE_LOCAL.
This is an artifact of the design when there was no PROCESS_LOCAL and
NODE_LOCAL was the best schedule possible (without explicitly having these
level : we had node and any).
TaskSchedulerImpl and TaskSetManager do not properly prioritize which tasks
get assigned to an executor
---
Key: SPARK-2294
URL: https://issues.apache.org/jira/browse/SPARK-2294
Project: Spark
Issue Type: Bug
Components: Spark Core
Affects Versions: 1.0.0, 1.0.1
Reporter: Kay Ousterhout
If an executor E is free, a task may be speculatively assigned to E when
there are other tasks in the job that have not been launched (at all) yet.
Similarly, a task without any locality preferences may be assigned to E when
there was another NODE_LOCAL task that could have been scheduled.
This happens because TaskSchedulerImpl calls TaskSetManager.resourceOffer
(which in turn calls TaskSetManager.findTask) with increasing locality
levels, beginning with PROCESS_LOCAL, followed by NODE_LOCAL, and so on until
the highest currently allowed level. Now, supposed NODE_LOCAL is the highest
currently allowed locality level. The first time findTask is called, it will
be called with max level PROCESS_LOCAL; if it cannot find any PROCESS_LOCAL
tasks, it will try to schedule tasks with no locality preferences or
speculative tasks. As a result, speculative tasks or tasks with no
preferences may be scheduled instead of NODE_LOCAL tasks.
cc [~matei]
--
This message was sent by Atlassian JIRA
(v6.2#6252)