[
https://issues.apache.org/jira/browse/SPARK-12554?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Lijie Xu updated SPARK-12554:
-----------------------------
Description:
In scheduleExecutorsOnWorker() in Master.scala,
*val keepScheduling = coresToAssign >= minCoresPerExecutor* should be changed
to *val keepScheduling = coresToAssign > 0*
Suppose that an app's requested cores is 10 (i.e., spark.cores.max = 10) and
app.coresPerExecutor is 4 (i.e., spark.executor.cores = 4).
After allocating two executors (each has 4 cores) to this app, the
*coresToAssign = 2* and *minCoresPerExecutor = coresPerExecutor = 4*, so
*keepScheduling = false* and no extra executor will be allocated to the app. If
*spark.scheduler.minRegisteredResourcesRatio* is set to a large number (e.g., >
0.8 in this case), the app will hang and never finish.
Another case: if a small app's coresPerExecutor is larger than its requested
cores (e.g., spark.cores.max = 10, spark.executor.cores = 16), *val
keepScheduling = coresToAssign >= minCoresPerExecutor* is always FALSE. As a
result, this app will hang and never finish.
was:
In scheduleExecutorsOnWorker() in Master.scala,
*val keepScheduling = coresToAssign >= minCoresPerExecutor* should be changed
to *val keepScheduling = coresToAssign > 0*
Suppose that an app's requested cores is 10 (i.e., spark.cores.max = 10) and
app.coresPerExecutor is 4 (i.e., spark.executor.cores = 4).
After allocating two executors (each has 4 cores) to this app, the
*coresToAssign = 2* and *minCoresPerExecutor = coresPerExecutor = 4*, so
*keepScheduling = false* and no extra executor will be allocated to the app. If
*spark.scheduler.minRegisteredResourcesRatio* is set to a large number (e.g., >
0.8 in this case), the app will hang and never finish.
In particular, if a small app's coresPerExecutor is larger than its requested
cores (e.g., spark.cores.max = 10, spark.executor.cores = 16), *val
keepScheduling = coresToAssign >= minCoresPerExecutor* is always FALSE. As a
result, this app will hang and never finish.
> Standalone app scheduler will hang when app.coreToAssign < minCoresPerExecutor
> ------------------------------------------------------------------------------
>
> Key: SPARK-12554
> URL: https://issues.apache.org/jira/browse/SPARK-12554
> Project: Spark
> Issue Type: Bug
> Components: Deploy, Scheduler
> Affects Versions: 1.5.2
> Reporter: Lijie Xu
> Priority: Critical
>
> In scheduleExecutorsOnWorker() in Master.scala,
> *val keepScheduling = coresToAssign >= minCoresPerExecutor* should be changed
> to *val keepScheduling = coresToAssign > 0*
> Suppose that an app's requested cores is 10 (i.e., spark.cores.max = 10) and
> app.coresPerExecutor is 4 (i.e., spark.executor.cores = 4).
> After allocating two executors (each has 4 cores) to this app, the
> *coresToAssign = 2* and *minCoresPerExecutor = coresPerExecutor = 4*, so
> *keepScheduling = false* and no extra executor will be allocated to the app.
> If *spark.scheduler.minRegisteredResourcesRatio* is set to a large number
> (e.g., > 0.8 in this case), the app will hang and never finish.
> Another case: if a small app's coresPerExecutor is larger than its requested
> cores (e.g., spark.cores.max = 10, spark.executor.cores = 16), *val
> keepScheduling = coresToAssign >= minCoresPerExecutor* is always FALSE. As a
> result, this app will hang and never finish.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
