flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sihua Zhou (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-7153) Eager Scheduling can't allocate source for ExecutionGraph correctly
Date Tue, 18 Jul 2017 13:58:00 GMT

    [ https://issues.apache.org/jira/browse/FLINK-7153?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16091575#comment-16091575
] 

Sihua Zhou commented on FLINK-7153:
-----------------------------------

After thinking about is, i found that this issue contains two subproblems. According to the
PR template's guidance, we should't mix them together. So i created two subtasks under this
issue.
1. [FLINK-7218|https://issues.apache.org/jira/browse/FLINK-7218] address the problem: ExecutionVertex.getPreferredLocationsBasedOnInputs()
always return empty.
2. [FLINK-7219|https://issues.apache.org/jira/browse/FLINK-7219] address the problem: Current
allocate strategy cann‘t achieve the optimal effect with input's location.

> Eager Scheduling can't allocate source for ExecutionGraph correctly
> -------------------------------------------------------------------
>
>                 Key: FLINK-7153
>                 URL: https://issues.apache.org/jira/browse/FLINK-7153
>             Project: Flink
>          Issue Type: Bug
>          Components: JobManager
>    Affects Versions: 1.3.1
>            Reporter: Sihua Zhou
>            Assignee: Stephan Ewen
>            Priority: Blocker
>             Fix For: 1.3.2
>
>
> The ExecutionGraph.scheduleEager() function allocate for ExecutionJobVertex one by one
via calling ExecutionJobVertex.allocateResourcesForAll(), here is two problem about it:
> 1. The ExecutionVertex.getPreferredLocationsBasedOnInputs will always return empty, cause
`sourceSlot` always be null until `ExectionVertex` has been deployed via 'Execution.deployToSlot()'.
So allocate resource base on prefered location can't work correctly, we need to set the slot
info for `Execution` as soon as Execution.allocateSlotForExecution() called successfully?
> 2. Current allocate strategy can't allocate the slot optimize.  Here is the test case:
> {code}
> JobVertex v1 = new JobVertex("v1", jid1);
> JobVertex v2 = new JobVertex("v2", jid2);
> SlotSharingGroup group = new SlotSharingGroup();
> v1.setSlotSharingGroup(group);
> v2.setSlotSharingGroup(group);
> v1.setParallelism(2);
> v2.setParallelism(4);
> v1.setInvokableClass(BatchTask.class);
> v2.setInvokableClass(BatchTask.class);
> v2.connectNewDataSetAsInput(v1, DistributionPattern.POINTWISE, ResultPartitionType.PIPELINED_BOUNDED);
> {code}
> Currently, after allocate for v1,v2, we got a local partition and three remote partition.
But actually, it should be 2 local partition and 2 remote partition. 
> The causes of the above problems is becuase that the current allocate strategy is allocate
the resource for execution one by one(if the execution can allocate from SlotGroup than get
it, Otherwise ask for a new one for it). 
> If we change the allocate strategy to two step will solve this problem, below is the
Pseudo code:
> {code}
> for (ExecutionJobVertex ejv: getVerticesTopologically) {
> //step 1: try to allocate from SlothGroup base on inputs one by one (which only allocate
resource base on location).
> //step 2: allocate for the remain execution.
> }
> {code}



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message