flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF GitHub Bot (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-6364) Implement incremental checkpointing in RocksDBStateBackend
Date Tue, 02 May 2017 17:33:06 GMT

    [ https://issues.apache.org/jira/browse/FLINK-6364?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15993347#comment-15993347

ASF GitHub Bot commented on FLINK-6364:

Github user StefanRRichter commented on a diff in the pull request:

    --- Diff: flink-contrib/flink-statebackend-rocksdb/src/main/java/org/apache/flink/contrib/streaming/state/RocksDBKeyedStateHandle.java
    @@ -0,0 +1,209 @@
    + * Licensed to the Apache Software Foundation (ASF) under one
    + * or more contributor license agreements.  See the NOTICE file
    + * distributed with this work for additional information
    + * regarding copyright ownership.  The ASF licenses this file
    + * to you under the Apache License, Version 2.0 (the
    + * "License"); you may not use this file except in compliance
    + * with the License.  You may obtain a copy of the License at
    + *
    + *     http://www.apache.org/licenses/LICENSE-2.0
    + *
    + * Unless required by applicable law or agreed to in writing, software
    + * distributed under the License is distributed on an "AS IS" BASIS,
    + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    + * See the License for the specific language governing permissions and
    + * limitations under the License.
    + */
    +package org.apache.flink.contrib.streaming.state;
    +import org.apache.flink.api.common.JobID;
    +import org.apache.flink.runtime.state.CompositeStateHandle;
    +import org.apache.flink.runtime.state.KeyGroupRange;
    +import org.apache.flink.runtime.state.KeyedStateHandle;
    +import org.apache.flink.runtime.state.SharedStateHandle;
    +import org.apache.flink.runtime.state.SharedStateRegistry;
    +import org.apache.flink.runtime.state.StateUtil;
    +import org.apache.flink.runtime.state.StreamStateHandle;
    +import org.apache.flink.util.Preconditions;
    +import org.slf4j.Logger;
    +import org.slf4j.LoggerFactory;
    +import java.util.Map;
    +import java.util.Set;
    + * The handle to states in incremental snapshots taken by {@link RocksDBKeyedStateBackend}
    + */
    +public class RocksDBKeyedStateHandle implements KeyedStateHandle, CompositeStateHandle
    +	private static final Logger LOG = LoggerFactory.getLogger(RocksDBKeyedStateHandle.class);
    +	private static final long serialVersionUID = -8328808513197388231L;
    +	private final JobID jobId;
    +	private final String operatorIdentifier;
    +	private final KeyGroupRange keyGroupRange;
    +	private final Set<String> newSstFileNames;
    --- End diff --
    Overall, I wonder how this class impacts larger deployments, where a lot of those state
handles are send via RPC. I suggest to keep the serialization footprint as small as possible.
For example, maybe we can eliminate `newSstFileNames` by introducing two `Map<String, StreamStateHandle>`:
`newSstFiles` and `previousSstFiles`, instead of checking against a set. This could even simplify
some methods, e.g. the registration or deletion code. We could then provide a method that
delivers an combined iterator over all `Entry<String, StreamStateHandle>`from both maps.

> Implement incremental checkpointing in RocksDBStateBackend
> ----------------------------------------------------------
>                 Key: FLINK-6364
>                 URL: https://issues.apache.org/jira/browse/FLINK-6364
>             Project: Flink
>          Issue Type: Sub-task
>          Components: State Backends, Checkpointing
>            Reporter: Xiaogang Shi
>            Assignee: Xiaogang Shi
> {{RocksDBStateBackend}} is well suited for incremental checkpointing because RocksDB
is base on LSM trees,  which record updates in new sst files and all sst files are immutable.
By only materializing those new sst files, we can significantly improve the performance of

This message was sent by Atlassian JIRA

View raw message