[
https://issues.apache.org/jira/browse/GOBBLIN-1688?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Matthew Ho updated GOBBLIN-1688:
--------------------------------
Description:
As part of the Helix Dynamic Workunit Allocation project, AM and TR need to be
able to communicate state of the currently running workunit s.t. the AM can
recompute the workunit into multiple smaller workunits when the task runner is
overloaded.
The goal of this ticket is to enable the AM to consume that message from the
FileSystem message buffer and act on the message by recomputing the workunit
into multiple smaller workunits
was:As part of the Helix Dynamic Workunit Allocation project, AM and TR need
to be able to communicate state of the currently running workunit s.t. the AM
can recompute the workunit into multiple smaller workunits when the task runner
is overloaded.
> Implementation splitting / recomputing workunit
> -----------------------------------------------
>
> Key: GOBBLIN-1688
> URL: https://issues.apache.org/jira/browse/GOBBLIN-1688
> Project: Apache Gobblin
> Issue Type: New Feature
> Reporter: Matthew Ho
> Priority: Major
>
> As part of the Helix Dynamic Workunit Allocation project, AM and TR need to
> be able to communicate state of the currently running workunit s.t. the AM
> can recompute the workunit into multiple smaller workunits when the task
> runner is overloaded.
> The goal of this ticket is to enable the AM to consume that message from the
> FileSystem message buffer and act on the message by recomputing the workunit
> into multiple smaller workunits
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
