Deprecated API

Contents

• Deprecated Fields
• Deprecated Methods
• Deprecated Enum Constants

Deprecated Fields

Field and Description
org.apache.hadoop.yarn.conf.YarnConfiguration.APPLICATION_MAX_TAG_LENGTH
org.apache.hadoop.yarn.conf.YarnConfiguration.APPLICATION_MAX_TAGS
org.apache.hadoop.yarn.conf.YarnConfiguration.AUTO_FAILOVER_EMBEDDED This property should never be set to false.
org.apache.hadoop.yarn.conf.YarnConfiguration.CURATOR_LEADER_ELECTOR Eventually, we want to default to the curator-based implementation and remove the ActiveStandbyElector based implementation. We should remove this config then.
org.apache.hadoop.yarn.conf.YarnConfiguration.DEFAULT_AUTO_FAILOVER_EMBEDDED The YarnConfiguration.AUTO_FAILOVER_EMBEDDED property is deprecated.
org.apache.hadoop.yarn.conf.YarnConfiguration.DEFAULT_NM_CONTAINER_MON_INTERVAL_MS
org.apache.hadoop.yarn.conf.YarnConfiguration.DEFAULT_NM_DOCKER_STOP_GRACE_PERIOD
org.apache.hadoop.yarn.conf.YarnConfiguration.DEFAULT_RM_STATE_STORE_MAX_COMPLETED_APPLICATIONS This default value is ignored and will be removed in a future release. The default value of yarn.resourcemanager.state-store.max-completed-applications is the value of YarnConfiguration.RM_MAX_COMPLETED_APPLICATIONS.
org.apache.hadoop.yarn.conf.YarnConfiguration.DISPLAY_APPS_FOR_LOGGED_IN_USER
org.apache.hadoop.yarn.conf.YarnConfiguration.HADOOP_HTTP_WEBAPP_SCHEDULER_PAGE This field is deprecated for YarnConfiguration.YARN_HTTP_WEBAPP_SCHEDULER_PAGE
org.apache.hadoop.yarn.conf.YarnConfiguration.NM_DOCKER_STOP_GRACE_PERIOD use YarnConfiguration.NM_SLEEP_DELAY_BEFORE_SIGKILL_MS
org.apache.hadoop.yarn.conf.YarnConfiguration.YARN_CLIENT_APP_SUBMISSION_POLL_INTERVAL_MS

Deprecated Methods

Method and Description
org.apache.hadoop.yarn.api.records.ApplicationSubmissionContext.getAMContainerResourceRequest() See ApplicationSubmissionContext.getAMContainerResourceRequests()
org.apache.hadoop.yarn.api.records.ContainerId.getId()
org.apache.hadoop.yarn.api.records.Resource.getMemory()
org.apache.hadoop.yarn.api.protocolrecords.GetClusterNodeLabelsResponse.getNodeLabels() Use GetClusterNodeLabelsResponse.getNodeLabelList() instead.
org.apache.hadoop.yarn.api.ContainerManagementProtocol.increaseContainersResource(IncreaseContainersResourceRequest)
org.apache.hadoop.yarn.api.records.ContainerId.newInstance(ApplicationAttemptId, int)
org.apache.hadoop.yarn.api.protocolrecords.GetClusterNodeLabelsResponse.newInstance(Set<String>) Use GetClusterNodeLabelsResponse.newInstance(List) instead.
org.apache.hadoop.yarn.api.records.ApplicationSubmissionContext.setAMContainerResourceRequest(ResourceRequest) See ApplicationSubmissionContext.setAMContainerResourceRequests(List)
org.apache.hadoop.yarn.api.records.Resource.setMemory(int)
org.apache.hadoop.yarn.api.protocolrecords.GetClusterNodeLabelsResponse.setNodeLabels(Set<String>) Use GetClusterNodeLabelsResponse.setNodeLabelList(List) instead.

Deprecated Enum Constants

Enum Constant and Description
org.apache.hadoop.yarn.api.records.AMCommand.AM_RESYNC Sent by Resource Manager when it is out of sync with the AM and wants the AM get back in sync. Note: Instead of sending this command, ApplicationMasterNotRegisteredException will be thrown when ApplicationMaster is out of sync with ResourceManager and ApplicationMaster is expected to re-register with RM by calling ApplicationMasterProtocol.registerApplicationMaster(RegisterApplicationMasterRequest)
org.apache.hadoop.yarn.api.records.AMCommand.AM_SHUTDOWN Sent by Resource Manager when it wants the AM to shutdown. Note: This command was earlier sent by ResourceManager to instruct AM to shutdown if RM had restarted. Now ApplicationAttemptNotFoundException will be thrown in case that RM has restarted and AM is supposed to handle this exception by shutting down itself.