[
https://issues.apache.org/jira/browse/FLINK-23402?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17383277#comment-17383277
]
Timo Walther commented on FLINK-23402:
--------------------------------------
Thanks for the feedback so far.
An update regarding the API: actually the decision was already made in FLIP-134:
{quote}
We also want to make the shuffle mode configurable because there might be cases
where a user wants BATCH execution mode but still wants the data to be
exchanged in a pipelined fashion. For this we propose a new configuration
option execution.shuffle-mode. We will not expose a dedicated setter in
StreamExecutionEnvironment, or ExecutionConfig
{quote}
Also, we cannot do {{RuntimeExecutionMode.BATCH.withAllExchangesBlocking()}}
because {{RuntimeExecutionMode}} is an enum and {{withAllExchangesBlocking}}
would then also show up the streaming mode. We can only check this during
runtime.
So we have two options: either we go with extended enum values such as
{{RuntimeExecutionMode.BATCH, RuntimeExecutionMode.BATCH_ALL_PIPELINED}} or as
the FLIP states with config options. I would go with the config option for now.
This also makes it easier for the Table API that exposes
{{EnvironmentSettings.inBatchMode()}} and can then simply rely on the config
option for advanced settings.
> Expose a consistent GlobalDataExchangeMode
> ------------------------------------------
>
> Key: FLINK-23402
> URL: https://issues.apache.org/jira/browse/FLINK-23402
> Project: Flink
> Issue Type: Sub-task
> Components: API / DataStream
> Reporter: Timo Walther
> Priority: Major
>
> The Table API makes the {{GlobalDataExchangeMode}} configurable via
> {{table.exec.shuffle-mode}}.
> In Table API batch mode the StreamGraph is configured with
> {{ALL_EDGES_BLOCKING}} and in DataStream API batch mode
> {{FORWARD_EDGES_PIPELINED}}.
> I would vote for unifying the exchange mode of both APIs so that complex SQL
> pipelines behave identical in {{StreamTableEnvironment}} and
> {{TableEnvironment}}. Also the feedback a got so far would make
> {{ALL_EDGES_BLOCKING}} a safer option to run pipelines successfully with
> limited resources.
> [~lzljs3620320]
> {quote}
> The previous history was like this:
> - The default value is pipeline, and we find that many times due to
> insufficient resources, the deployment will hang. And the typical use of
> batch jobs is small resources running large parallelisms, because in batch
> jobs, the granularity of failover is related to the amount of data processed
> by a single task. The smaller the amount of data, the faster the fault
> tolerance. So most of the scenarios are run with small resources and large
> parallelisms, little by little slowly running.
> - Later, we switched the default value to blocking. We found that the better
> blocking shuffle implementation would not slow down the running speed much.
> We tested tpc-ds and it took almost the same time.
> {quote}
> [~dwysakowicz]
> {quote}
> I don't see a problem with changing the default value for DataStream batch
> mode if you think ALL_EDGES_BLOCKING is the better default option.
> {quote}
> In any case, we should make this configurable for DataStream API users and
> make the specific Table API option obsolete.
> It would include the following steps:
> - Move {{GlobalDataExchangeMode}} from {{o.a.f.streaming.api.graph}} to
> {{o.a.f.api.common}} (with reworked JavaDocs) as {{ExchangeMode}} (to have it
> shorter) next to {{RuntimeMode}}
> - Add {{StreamExecutionEnvironment.setExchangeMode()}} next to
> {{setRuntimeMode}}
> - Add option {{execution.exchange-mode}}
> - Add checks for invalid combinations to StreamGraphGenerator
> - Deprecate ExecutionMode ([avoid
> confusion|https://stackoverflow.com/questions/68335472/what-is-difference-in-runtimeexecutionmode-and-executionmode])
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
