slfan1989 commented on code in PR #4846:
URL: https://github.com/apache/hadoop/pull/4846#discussion_r964252561


##########
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-router/src/main/java/org/apache/hadoop/yarn/server/router/clientrm/FederationClientInterceptor.java:
##########
@@ -949,7 +949,7 @@ public ReservationSubmissionResponse submitReservation(
         // Second, determine whether the current ReservationId has a 
corresponding subCluster.
         // If it does not exist, add it. If it exists, update it.
         Boolean exists = existsReservationHomeSubCluster(reservationId);
-        if (!exists) {
+        if (!exists || i == 0) {

Review Comment:
   Reservation may be repeatedly submitted (the same reservationId is submitted 
repeatedly). At this time, we should try to use the first reservation result. 
If the first reservation encounters some abnormalities, find another RM for 
reservation.
   
   If a reservationId(**reservationId_1**) has successfully applied for 
reserved resources in SC_1, and the **reservationId_1** is submitted 
repeatedly, we should try to use the allocation result of SC_1
   
   For the case of **reservationId_1**, the following code will return true, 
   ```
   Boolean exists = existsReservationHomeSubCluster(reservationId);
   ```
   
   Then the first subcluster(SC_1) will be used instead of SC2
   ```
   updateReservationHomeSubCluster(subClusterId, reservationId, 
reservationHomeSubCluster);
   ```
   
   But at this time we still want to go 
`addReservationHomeSubCluster(reservationId, reservationHomeSubCluster)`
   `addReservationHomeSubCluster`, will not be added repeatedly for the 
existing reservationId.
   
   I added a condition, if i=0, continue to use `addReservationHomeSubCluster`
   
   



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: common-issues-unsubscr...@hadoop.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: common-issues-h...@hadoop.apache.org

Reply via email to