[ 
https://issues.apache.org/jira/browse/FLINK-9499?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16547783#comment-16547783
 ] 

ASF GitHub Bot commented on FLINK-9499:
---------------------------------------

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

    https://github.com/apache/flink/pull/6330#discussion_r203365784
  
    --- Diff: 
flink-runtime-web/src/main/java/org/apache/flink/runtime/webmonitor/handlers/JarRunHandler.java
 ---
    @@ -138,12 +154,22 @@ public JarRunHandler(
                        });
        }
     
    -   private static SavepointRestoreSettings getSavepointRestoreSettings(
    -                   final @Nonnull HandlerRequest<EmptyRequestBody, 
JarRunMessageParameters> request)
    +   private SavepointRestoreSettings getSavepointRestoreSettings(
    +                   final @Nonnull HandlerRequest<JarRunRequestBody, 
JarRunMessageParameters> request)
                                throws RestHandlerException {
     
    -           final boolean allowNonRestoredState = 
getQueryParameter(request, AllowNonRestoredStateQueryParameter.class, false);
    -           final String savepointPath = getQueryParameter(request, 
SavepointPathQueryParameter.class);
    +           final JarRunRequestBody requestBody = request.getRequestBody();
    +
    +           final boolean allowNonRestoredState = 
fromRequestBodyOrQueryParameter(
    +                   requestBody.getAllowNonRestoredState(),
    +                   () -> getQueryParameter(request, 
AllowNonRestoredStateQueryParameter.class),
    +                   false,
    +                   log);
    +           final String savepointPath = fromRequestBodyOrQueryParameter(
    --- End diff --
    
    True, making the specification of the default value explicit will better 
guard against this. In the other case, the user would have to know the 
difference between these methods and which to apply to primitive and 
non-primitive values. Alright, then it's good to go from my side.


> Allow REST API for running a job to provide job configuration as body of POST 
> request
> -------------------------------------------------------------------------------------
>
>                 Key: FLINK-9499
>                 URL: https://issues.apache.org/jira/browse/FLINK-9499
>             Project: Flink
>          Issue Type: Improvement
>          Components: REST
>    Affects Versions: 1.3.2
>            Reporter: Esteban Serrano
>            Assignee: Esteban Serrano
>            Priority: Minor
>              Labels: pull-request-available
>
> Based on 
> [this|https://ci.apache.org/projects/flink/flink-docs-release-1.3/monitoring/rest_api.html#submitting-programs]
>  documentation, the REST API provides a way to submit a request for running a 
> Flink job. The POST request must include the job configuration information as 
> query parameters using the documented parameter names ("program-args", 
> "entry-class", "parallelism", etc.) 
> Depending on the job parameters, the full URL for the POST request can reach 
> a size that is over the maximum size (currently at 4096 bytes) of what is 
> allowed by the configuration of Netty. To overcome this, it would be useful 
> to allow users to provide the job configuration not only as query parameters 
> but also as POST parameters. 
> For the most part, it is the "program-args" parameter that can make the URL 
> grow in size based on the needs of the developer and the job. All other 
> attributes should be pretty constant.  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to