Pergunta
In Hadoop 2.2, we can use the High Availability feature to deal with the namenode's fault. Is there any solutions to deal with the ResourceManager's fault? If the machine which has the ResourceManager's process is shutdown or offline, could any other machine backup the states. Thanks.
Solução
In YARN, when a ResourceManager fails over, all applications will be restarted. For applications like Hadoop MapReduce which implemented recovery, jobs will continue from their last saved progress.
Failover handling of ResourceManager will be implemented in the future.
Outras dicas
Resource Manager HA developing status is tracked here, https://issues.apache.org/jira/browse/YARN-149. And work preserving restart is tracked here, https://issues.apache.org/jira/browse/YARN-556. Both of them are under development
https://stackoverflow.com/questions/22501103
italiano
english
français
española
中国
日本の
العربية
Deutsch
한국어
Português
Russian