Skip to content

[SPARK-15606][core] Use non-blocking removeExecutor call to avoid deadlocks #13355

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 7 commits into from
Closed

[SPARK-15606][core] Use non-blocking removeExecutor call to avoid deadlocks #13355

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
139e875
Use a minimum of 3 dispatcher threads to avoid deadlocks
robbinspg May 27, 2016
c991d70
Don't block in BlockManagerMaster.removeExecutor
robbinspg May 31, 2016
6a95c50
Create new remonveExecutorAsync method
robbinspg Jun 1, 2016
eb900d8
scalastyle fix
robbinspg Jun 1, 2016
68be9d8
Merge branch 'master' of https://github.com/apache/spark.git
robbinspg Jun 1, 2016
180a361
Merge branch 'master' of https://github.com/apache/spark.git
robbinspg Jun 1, 2016
4d3fc6f
remove slow test suite
robbinspg Jun 1, 2016
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion core/src/main/scala/org/apache/spark/scheduler/cluster/CoarseGrainedSchedulerBackend.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -295,7 +295,7 @@ class CoarseGrainedSchedulerBackend(scheduler: TaskSchedulerImpl, val rpcEnv: Rp
// manager to reregister itself. If that happens, the block manager master will know
// about the executor, but the scheduler will not. Therefore, we should remove the
// executor from the block manager when we hit this case.
scheduler.sc.env.blockManager.master.removeExecutor(executorId)
scheduler.sc.env.blockManager.master.removeExecutorAsync(executorId)
logInfo(s"Asked to remove non-existent executor $executorId")
}
}
Expand Down
8 changes: 8 additions & 0 deletions core/src/main/scala/org/apache/spark/storage/BlockManagerMaster.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -42,6 +42,14 @@ class BlockManagerMaster(
logInfo("Removed " + execId + " successfully in removeExecutor")
}

/** Request removal of a dead executor from the driver endpoint.
* This is only called on the driver side. Non-blocking
*/
def removeExecutorAsync(execId: String) {
driverEndpoint.ask[Boolean](RemoveExecutor(execId))
logInfo("Removal of executor " + execId + " requested")
}

/** Register the BlockManager's id with the driver. */
def registerBlockManager(
blockManagerId: BlockManagerId, maxMemSize: Long, slaveEndpoint: RpcEndpointRef): Unit = {
Expand Down