Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

Currently, the different artifacts are cleaned up in different locations:

Component

Who is responsible?

Why?

JobResultStore

Dispatcher

It needs to work without a JobMaster.

JobGraphStore

Dispatcher

We need to be able to fetch a JobGraph for recovery before starting JobMaster.

HighAvailableServices

  • JM LeaderElection
  • RM LeaderElection

Dispatcher

We can clean this up after there won’t be any more contenders.

BlobServer (for a given job)

Dispatcher


CheckpointCounter/CheckpointStore 

JobMaster

This is tied to a single job execution (JobMasterLifecycle).

JobMaster

Dispatcher

Dispatcher is responsible for JobMaster management.

The plan is to provide a common interface that enables all components to trigger the cleanup based on the JobID. This cleanup component can encapsulate the retry mechanism. The actual cleanup can then be triggered either in the JobMaster or the Dispatcher depending on whether a JobGraph is still available or not.

...

draw.io Diagram
bordertrue
diagramNameFLIP-140 194 - Interfaces
simpleViewerfalse
width
linksauto
tbstyletop
lboxtrue
diagramWidth561
revision2

...