xintongsong commented on a change in pull request #12444: URL: https://github.com/apache/flink/pull/12444#discussion_r436315438
########## File path: flink-yarn/src/main/java/org/apache/flink/yarn/YarnClusterDescriptor.java ########## @@ -508,7 +508,11 @@ public void killCluster(ApplicationId applicationId) throws FlinkException { throw new YarnDeploymentException("Could not retrieve information about free cluster resources.", e); } - final int yarnMinAllocationMB = yarnConfiguration.getInt(YarnConfiguration.RM_SCHEDULER_MINIMUM_ALLOCATION_MB, 0); + final int yarnMinAllocationMB = yarnConfiguration.getInt( + YarnConfiguration.RM_SCHEDULER_MINIMUM_ALLOCATION_MB, + YarnConfiguration.DEFAULT_RM_SCHEDULER_MINIMUM_ALLOCATION_MB); + Preconditions.checkArgument(yarnMinAllocationMB > 0, "The minimum allocation memory " + + "(" + yarnMinAllocationMB + " MB) configured via 'yarn.scheduler.minimum-allocation-mb' should be greater than 0."); Review comment: The string literal can be replaced with `YarnConfiguration.RM_SCHEDULER_MINIMUM_ALLOCATION_MB`. ########## File path: flink-yarn/src/main/java/org/apache/flink/yarn/YarnClusterDescriptor.java ########## @@ -508,7 +508,9 @@ public void killCluster(ApplicationId applicationId) throws FlinkException { throw new YarnDeploymentException("Could not retrieve information about free cluster resources.", e); } - final int yarnMinAllocationMB = yarnConfiguration.getInt(YarnConfiguration.RM_SCHEDULER_MINIMUM_ALLOCATION_MB, 0); + final int yarnMinAllocationMB = yarnConfiguration.getInt( Review comment: I think it would be good to always try to throw an exception with the most descriptive type if possible. IMO, this should have higher priority than keeping the codes concise. In this case, I would suggest to throw either a `YarnDeploymentException` or a `IllegalConfigurationException`, rather than `IllegalStateException` that `Preconditions.checkArgument` throws. ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org