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-3231) Handle Kinesis-side resharding in Kinesis streaming consumer
Date Fri, 01 Jul 2016 09:45:11 GMT

    [ https://issues.apache.org/jira/browse/FLINK-3231?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15358714#comment-15358714
] 

ASF GitHub Bot commented on FLINK-3231:
---------------------------------------

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

    https://github.com/apache/flink/pull/2131#discussion_r69274851
  
    --- Diff: docs/apis/streaming/connectors/kinesis.md ---
    @@ -60,10 +60,10 @@ to setup Kinesis streams. Make sure to create the appropriate IAM
policy and use
     
     ### Kinesis Consumer
    --- End diff --
    
    I think we should also mention the threading model of the Kinesis consumer in the documentation.
    
    Users should know that each parallel Flink instance will constantly query Kinesis for
shards. So if a user has 5 shards, but runs Flink with a parallelism of 50, there will be
50 threads constantly querying Kinesis.
    
    Also, we should explain that there will always be one thread per shard running.


> Handle Kinesis-side resharding in Kinesis streaming consumer
> ------------------------------------------------------------
>
>                 Key: FLINK-3231
>                 URL: https://issues.apache.org/jira/browse/FLINK-3231
>             Project: Flink
>          Issue Type: Sub-task
>          Components: Kinesis Connector, Streaming Connectors
>    Affects Versions: 1.1.0
>            Reporter: Tzu-Li (Gordon) Tai
>            Assignee: Tzu-Li (Gordon) Tai
>             Fix For: 1.1.0
>
>
> A big difference between Kinesis shards and Kafka partitions is that Kinesis users can
choose to "merge" and "split" shards at any time for adjustable stream throughput capacity.
This article explains this quite clearly: https://brandur.org/kinesis-by-example.
> This will break the static shard-to-task mapping implemented in the basic version of
the Kinesis consumer (https://issues.apache.org/jira/browse/FLINK-3229). The static shard-to-task
mapping is done in a simple round-robin-like distribution which can be locally determined
at each Flink consumer task (Flink Kafka consumer does this too).
> To handle Kinesis resharding, we will need some way to let the Flink consumer tasks coordinate
which shards they are currently handling, and allow the tasks to ask the coordinator for a
shards reassignment when the task finds out it has found a closed shard at runtime (shards
will be closed by Kinesis when it is merged and split).
> We need a centralized coordinator state store which is visible to all Flink consumer
tasks. Tasks can use this state store to locally determine what shards it can be reassigned.
Amazon KCL uses a DynamoDB table for the coordination, but as described in https://issues.apache.org/jira/browse/FLINK-3211,
we unfortunately can't use KCL for the implementation of the consumer if we want to leverage
Flink's checkpointing mechanics. For our own implementation, Zookeeper can be used for this
state store, but that means it would require the user to set up ZK to work.
> Since this feature introduces extensive work, it is opened as a separate sub-task from
the basic implementation https://issues.apache.org/jira/browse/FLINK-3229.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message