big-data/hadoop
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

YARN-8860. Federation client intercepter class contains unwanted character. Contributed by Abhishek Modi.

Browse Source
This commit is contained in:
Giovanni Matteo Fumarola 2018-11-16 10:23:06 -08:00
parent 3420ddd0f8
commit abd6d48c46
2 changed files with 60 additions and 60 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

60
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-router/src/main/java/org/apache/hadoop/yarn/server/router/clientrm/FederationClientInterceptor.java
View File

@ -304,19 +304,19 @@ public GetNewApplicationResponse getNewApplication(
*
* Base scenarios:
*
* The Client submits an application to the Router. • The Router selects one
* SubCluster to forward the request. • The Router inserts a tuple into
* StateStore with the selected SubCluster (e.g. SC1) and the appId. • The
* State Store replies with the selected SubCluster (e.g. SC1). • The Router
* The Client submits an application to the Router. The Router selects one
* SubCluster to forward the request. The Router inserts a tuple into
* StateStore with the selected SubCluster (e.g. SC1) and the appId. The
* State Store replies with the selected SubCluster (e.g. SC1). The Router
* submits the request to the selected SubCluster.
*
* In case of State Store failure:
*
* The client submits an application to the Router. • The Router selects one
* SubCluster to forward the request. • The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC1) and the appId. • Due to the
* The client submits an application to the Router. The Router selects one
* SubCluster to forward the request. The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC1) and the appId. Due to the
* State Store down the Router times out and it will retry depending on the
* FederationFacade settings. • The Router replies to the client with an error
* FederationFacade settings. The Router replies to the client with an error
* message.
*
* If State Store fails after inserting the tuple: identical behavior as
@ -326,26 +326,26 @@ public GetNewApplicationResponse getNewApplication(
*
* Scenario 1 – Crash before submission to the ResourceManager
*
* The Client submits an application to the Router. • The Router selects one
* SubCluster to forward the request. • The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC1) and the appId. • The Router
* crashes. • The Client timeouts and resubmits the application. • The Router
* selects one SubCluster to forward the request. • The Router inserts a tuple
* into State Store with the selected SubCluster (e.g. SC2) and the appId. •
* The Client submits an application to the Router. The Router selects one
* SubCluster to forward the request. The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC1) and the appId. The Router
* crashes. The Client timeouts and resubmits the application. The Router
* selects one SubCluster to forward the request. The Router inserts a tuple
* into State Store with the selected SubCluster (e.g. SC2) and the appId.
* Because the tuple is already inserted in the State Store, it returns the
* previous selected SubCluster (e.g. SC1). • The Router submits the request
* previous selected SubCluster (e.g. SC1). The Router submits the request
* to the selected SubCluster (e.g. SC1).
*
* Scenario 2 – Crash after submission to the ResourceManager
*
* • The Client submits an application to the Router. • The Router selects one
* SubCluster to forward the request. • The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC1) and the appId. • The Router
* submits the request to the selected SubCluster. • The Router crashes. • The
* Client timeouts and resubmit the application. • The Router selects one
* SubCluster to forward the request. • The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC2) and the appId. • The State
* Store replies with the selected SubCluster (e.g. SC1). • The Router submits
* The Client submits an application to the Router. The Router selects one
* SubCluster to forward the request. The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC1) and the appId. The Router
* submits the request to the selected SubCluster. The Router crashes. The
* Client timeouts and resubmit the application. The Router selects one
* SubCluster to forward the request. The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC2) and the appId. The State
* Store replies with the selected SubCluster (e.g. SC1). The Router submits
* the request to the selected SubCluster (e.g. SC1). When a client re-submits
* the same application to the same RM, it does not raise an exception and
* replies with operation successful message.
@ -354,14 +354,14 @@ public GetNewApplicationResponse getNewApplication(
*
* In case of ResourceManager failure:
*
* The Client submits an application to the Router. • The Router selects one
* SubCluster to forward the request. • The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC1) and the appId. • The Router
* submits the request to the selected SubCluster. • The entire SubCluster is
* down – all the RMs in HA or the master RM is not reachable. • The Router
* times out. • The Router selects a new SubCluster to forward the request. •
* The Client submits an application to the Router. The Router selects one
* SubCluster to forward the request. The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC1) and the appId. The Router
* submits the request to the selected SubCluster. The entire SubCluster is
* down – all the RMs in HA or the master RM is not reachable. The Router
* times out. The Router selects a new SubCluster to forward the request.
* The Router update a tuple into State Store with the selected SubCluster
* (e.g. SC2) and the appId. • The State Store replies with OK answer. • The
* (e.g. SC2) and the appId. The State Store replies with OK answer. The
* Router submits the request to the selected SubCluster (e.g. SC2).
*/
@Override

60
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-router/src/main/java/org/apache/hadoop/yarn/server/router/webapp/FederationInterceptorREST.java
View File

@ -325,19 +325,19 @@ public Response createNewApplication(HttpServletRequest hsr)
* <p>
* Base scenarios:
* <p>
* The Client submits an application to the Router. • The Router selects one
* SubCluster to forward the request. • The Router inserts a tuple into
* StateStore with the selected SubCluster (e.g. SC1) and the appId. • The
* State Store replies with the selected SubCluster (e.g. SC1). • The Router
* The Client submits an application to the Router. The Router selects one
* SubCluster to forward the request. The Router inserts a tuple into
* StateStore with the selected SubCluster (e.g. SC1) and the appId. The
* State Store replies with the selected SubCluster (e.g. SC1). The Router
* submits the request to the selected SubCluster.
* <p>
* In case of State Store failure:
* <p>
* The client submits an application to the Router. • The Router selects one
* SubCluster to forward the request. • The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC1) and the appId. • Due to the
* The client submits an application to the Router. The Router selects one
* SubCluster to forward the request. The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC1) and the appId. Due to the
* State Store down the Router times out and it will retry depending on the
* FederationFacade settings. • The Router replies to the client with an error
* FederationFacade settings. The Router replies to the client with an error
* message.
* <p>
* If State Store fails after inserting the tuple: identical behavior as
@ -347,26 +347,26 @@ public Response createNewApplication(HttpServletRequest hsr)
* <p>
* Scenario 1 – Crash before submission to the ResourceManager
* <p>
* The Client submits an application to the Router. • The Router selects one
* SubCluster to forward the request. • The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC1) and the appId. • The Router
* crashes. • The Client timeouts and resubmits the application. • The Router
* selects one SubCluster to forward the request. • The Router inserts a tuple
* into State Store with the selected SubCluster (e.g. SC2) and the appId. •
* The Client submits an application to the Router. The Router selects one
* SubCluster to forward the request. The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC1) and the appId. The Router
* crashes. The Client timeouts and resubmits the application. The Router
* selects one SubCluster to forward the request. The Router inserts a tuple
* into State Store with the selected SubCluster (e.g. SC2) and the appId.
* Because the tuple is already inserted in the State Store, it returns the
* previous selected SubCluster (e.g. SC1). • The Router submits the request
* previous selected SubCluster (e.g. SC1). The Router submits the request
* to the selected SubCluster (e.g. SC1).
* <p>
* Scenario 2 – Crash after submission to the ResourceManager
* <p>
* • The Client submits an application to the Router. • The Router selects one
* SubCluster to forward the request. • The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC1) and the appId. • The Router
* submits the request to the selected SubCluster. • The Router crashes. • The
* Client timeouts and resubmit the application. • The Router selects one
* SubCluster to forward the request. • The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC2) and the appId. • The State
* Store replies with the selected SubCluster (e.g. SC1). • The Router submits
* The Client submits an application to the Router. The Router selects one
* SubCluster to forward the request. The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC1) and the appId. The Router
* submits the request to the selected SubCluster. The Router crashes. The
* Client timeouts and resubmit the application. The Router selects one
* SubCluster to forward the request. The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC2) and the appId. The State
* Store replies with the selected SubCluster (e.g. SC1). The Router submits
* the request to the selected SubCluster (e.g. SC1). When a client re-submits
* the same application to the same RM, it does not raise an exception and
* replies with operation successful message.
@ -375,14 +375,14 @@ public Response createNewApplication(HttpServletRequest hsr)
* <p>
* In case of ResourceManager failure:
* <p>
* The Client submits an application to the Router. • The Router selects one
* SubCluster to forward the request. • The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC1) and the appId. • The Router
* submits the request to the selected SubCluster. • The entire SubCluster is
* down – all the RMs in HA or the master RM is not reachable. • The Router
* times out. • The Router selects a new SubCluster to forward the request. •
* The Client submits an application to the Router. The Router selects one
* SubCluster to forward the request. The Router inserts a tuple into State
* Store with the selected SubCluster (e.g. SC1) and the appId. The Router
* submits the request to the selected SubCluster. The entire SubCluster is
* down – all the RMs in HA or the master RM is not reachable. The Router
* times out. The Router selects a new SubCluster to forward the request.
* The Router update a tuple into State Store with the selected SubCluster
* (e.g. SC2) and the appId. • The State Store replies with OK answer. • The
* (e.g. SC2) and the appId. The State Store replies with OK answer. The
* Router submits the request to the selected SubCluster (e.g. SC2).
*/
@Override