Hi Pankaj, First, you need to prepare a hadoop environment separately, including hdfs and Yarn. If you are familiar with hadoop, you could download the binary[1] and start the cluster on you nodes manually. Otherwise, some tools may help you to deploy a hadoop cluster, ambari[2] and cloudera manager[2].
Then, download the Flink with Hadoop pre-bundle. You can submit your flink job now. Please make sure you set the correct HADOOP_CONF_DIR in your flink client before starting a flink cluster. [1].https://hadoop.apache.org/releases.html [2]. https://www.cloudera.com/products/open-source/apache-hadoop/apache-ambari.html [3]. https://www.cloudera.com/products/product-components/cloudera-manager.html Ana <otakunon...@gmail.com> 于2019年11月20日周三 上午10:12写道: > Hi, > > I was able to run Flink on YARN by installing YARN and Flink separately. > > Thank you. > > Ana > > On Wed, Nov 20, 2019 at 10:42 AM Pankaj Chand <pankajchanda...@gmail.com> > wrote: > >> Hello, >> >> I want to run Flink on YARN upon a cluster of nodes. From the >> documentation, I was not able to fully understand how to go about it. Some >> of the archived answers are a bit old and had pending JIRA issues, so I >> thought I would ask. >> >> Am I first supposed to install YARN separately, and then download the >> Flink file and Hadoop pre-bundle? Or does the Hadoop-prebundle that I put >> into Flink's /lib folder provide the entire YARN installation? >> >> Is there any download that bundles a complete *installation of Fink + >> installation of YARN*? >> >> Thank you, >> >> Pankaj >> >