Interface TierMasterAgent

    • Method Detail

      • unregisterJob

        void unregisterJob​(org.apache.flink.api.common.JobID jobID)
        Unregister a job id.
      • restoreState

        default void restoreState​(TieredShuffleMasterSnapshot snapshot,
                                  org.apache.flink.api.common.JobID jobId)
        Restores the state of the tier master agent from the provided snapshots.
      • restoreState

        default void restoreState​(TieredShuffleMasterSnapshot snapshot)
        Restores the state of the tier master agent from the provided snapshots for the specified job.
      • getPartitionWithMetrics

        default CompletableFuture<Map<ResultPartitionID,​ShuffleMetrics>> getPartitionWithMetrics​(org.apache.flink.api.common.JobID jobId,
                                                                                                       Duration timeout,
                                                                                                       Set<ResultPartitionID> expectedPartitions)
        Retrieves specified partitions and their metrics (identified by expectedPartitions), the metrics include sizes of sub-partitions in a result partition.
        Parameters:
        jobId - ID of the target job
        timeout - The timeout used for retrieve the specified partitions.
        expectedPartitions - The set of identifiers for the result partitions whose metrics are to be fetched.
        Returns:
        A future will contain a map of the partitions with their metrics that could be retrieved from the expected partitions within the specified timeout period.
      • releasePartition

        void releasePartition​(TierShuffleDescriptor shuffleDescriptor)
        Release a tiered storage partition.
        Parameters:
        shuffleDescriptor - the partition shuffle descriptor to be released
      • close

        void close()
        Close this tier master agent.
      • partitionInRemote

        default boolean partitionInRemote()
        Is this tier manage the partition in remote cluster instead of flink taskmanager.