Deploy Resource Manager HA blueprint

jon_udaondo · ‎05-09-2018

Hello I have the following files and I want to deploy an HA ResourceManager:

cluster_configuration.json

{
  "Blueprints": {
    "stack_name": "HDP",
    "stack_version": "2.6"
  },
  "host_groups": [
    {
      "name": "services",
      "cardinality" : "1",
      "components": [
        { "name" : "HST_AGENT" },
        { "name" : "MAPREDUCE2_CLIENT" },
        { "name" : "YARN_CLIENT" },
        { "name" : "HDFS_CLIENT" },
        { "name" : "HBASE_CLIENT" },
        { "name" : "METRICS_MONITOR" },
        { "name" : "ZOOKEEPER_CLIENT" }
      ]
    },
    {
      "name": "datanodes",
      "cardinality" : "3",
      "components": [
        { "name" : "NODEMANAGER" },
        { "name" : "HST_AGENT" },
        { "name" : "MAPREDUCE2_CLIENT" },
        { "name" : "YARN_CLIENT" },
        { "name" : "HDFS_CLIENT" },
        { "name" : "HBASE_REGIONSERVER" },
        { "name" : "DATANODE" },
        { "name" : "METRICS_MONITOR" },
        { "name" : "ZOOKEEPER_CLIENT" },
        { "name" : "HBASE_CLIENT" }
      ]
    },
    {
      "name": "namenode1",
      "cardinality" : "1",
      "components": [
        { "name" : "HST_AGENT" },
        { "name" : "HDFS_CLIENT" },
        { "name" : "ZKFC" },
        { "name" : "ZOOKEEPER_SERVER" },
        { "name" : "HST_SERVER" },
        { "name" : "HBASE_CLIENT" },
        { "name" : "METRICS_MONITOR" },
        { "name" : "JOURNALNODE" },
        { "name" : "HBASE_MASTER" },
        { "name" : "NAMENODE" },
        { "name" : "APP_TIMELINE_SERVER" },
        { "name" : "METRICS_GRAFANA" }
      ]
    },
    {
      "name": "namenode2",
      "cardinality" : "1",
      "components": [
        { "name" : "HDFS_CLIENT" },
        { "name" : "ZKFC" },
        { "name" : "HISTORYSERVER" },
        { "name" : "METRICS_MONITOR" },
        { "name" : "HBASE_MASTER" },
        { "name" : "NAMENODE" },
        { "name" : "ACTIVITY_EXPLORER" },
        { "name" : "HST_AGENT" },
        { "name" : "ZOOKEEPER_SERVER" },
        { "name" : "HBASE_CLIENT" },
        { "name" : "JOURNALNODE" },
        { "name" : "METRICS_COLLECTOR" },
        { "name" : "RESOURCEMANAGER" }
      ]
    },
    {
      "name": "namenode3",
      "cardinality" : "1",
      "components": [
        { "name" : "ACTIVITY_ANALYZER" },
        { "name" : "HST_AGENT" },
        { "name" : "MAPREDUCE2_CLIENT" },
        { "name" : "YARN_CLIENT" },
        { "name" : "HDFS_CLIENT" },
        { "name" : "ZOOKEEPER_SERVER" },
        { "name" : "HBASE_CLIENT" },
        { "name" : "METRICS_MONITOR" },
        { "name" : "JOURNALNODE" },
        { "name" : "RESOURCEMANAGER" }
      ]
    }
  ],
  "configurations": [
    {
      "core-site": {
        "properties" : {
          "fs.defaultFS" : "hdfs://HACluster",
          "ha.zookeeper.quorum": "%HOSTGROUP::namenode1%:2181,%HOSTGROUP::namenode3%:2181",
          "hadoop.proxyuser.yarn.hosts": "%HOSTGROUP::namenode2%,%HOSTGROUP::namenode3%"
        }
      }
    },
    { "hdfs-site": {
        "properties" : {
          "dfs.client.failover.proxy.provider.HACluster" : "org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider",
          "dfs.ha.automatic-failover.enabled" : "true",
          "dfs.ha.fencing.methods" : "shell(/bin/true)",
          "dfs.ha.namenodes.HACluster" : "nn1,nn2",
          "dfs.namenode.http-address" : "%HOSTGROUP::namenode1%:50070",
          "dfs.namenode.http-address.HACluster.nn1" : "%HOSTGROUP::namenode1%:50070",
          "dfs.namenode.http-address.HACluster.nn2" : "%HOSTGROUP::namenode2%:50070",
          "dfs.namenode.https-address" : "%HOSTGROUP::namenode1%:50470",
          "dfs.namenode.https-address.HACluster.nn1" : "%HOSTGROUP::namenode1%:50470",
          "dfs.namenode.https-address.HACluster.nn2" : "%HOSTGROUP::namenode2%:50470",
          "dfs.namenode.rpc-address.HACluster.nn1" : "%HOSTGROUP::namenode1%:8020",
          "dfs.namenode.rpc-address.HACluster.nn2" : "%HOSTGROUP::namenode2%:8020",
          "dfs.namenode.shared.edits.dir" : "qjournal://%HOSTGROUP::namenode1%:8485;%HOSTGROUP::namenode2%:8485;%HOSTGROUP::namenode3%:8485/HACluster",
          "dfs.nameservices" : "HACluster"
        }
      }
    },
    { "yarn-site": {
        "properties": {
          "yarn.resourcemanager.ha.enabled": "true",
          "yarn.resourcemanager.ha.rm-ids": "rm1,rm2",
          "yarn.resourcemanager.hostname.rm1": "%HOSTGROUP::namenode3%",
          "yarn.resourcemanager.hostname.rm2": "%HOSTGROUP::namenode2%",
          "yarn.resourcemanager.webapp.address.rm1": "%HOSTGROUP::namenode3%:8088",
          "yarn.resourcemanager.webapp.address.rm2": "%HOSTGROUP::namenode2%:8088",
          "yarn.resourcemanager.webapp.https.address.rm1": "%HOSTGROUP::namenode3%:8090",
          "yarn.resourcemanager.webapp.https.address.rm2": "%HOSTGROUP::namenode2%:8090",
          "yarn.resourcemanager.recovery.enabled": "true",
          "yarn.resourcemanager.store.class": "org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore",
          "yarn.resourcemanager.zk-address": "%HOSTGROUP::namenode1%:2181,%HOSTGROUP::namenode3%:2181,%HOSTGROUP::namenode2%:2181",
          "yarn.client.failover-proxy-provider": "org.apache.hadoop.yarn.client.ConfiguredRMFailoverProxyProvider",
          "yarn.resourcemanager.cluster-id": "yarn-cluster",
          "yarn.resourcemanager.ha.automatic-failover.zk-base-path": "/yarn-leader-election"
        }
      }
    },
    {
      "hdfs-site" : {
        "properties_attributes" : { },
          "properties" : {
            "dfs.datanode.data.dir" : "/mnt/secondary1,/mnt/secondary2"
          }
      }
    },
    {
      "hadoop-env" : {
        "properties_attributes" : { },
          "properties" : {
            "namenode_heapsize" : "2048m"
          }
      }
    },
    {
      "activity-zeppelin-shiro": {
        "properties": {
          "users.admin": "admin"
        }
      }
    },
    {
      "hbase-site" : {
        "properties" : {
          "hbase.rootdir" : "hdfs://HACluster/apps/hbase/data"
        }
      }
   },
   {
     "hbase-env" : {
       "properties" : {
         "hbase_regionserver_heapsize" : "1024"
       }
     }
  }
  ]
}

hdputils-repo.json

{
   "Repositories":{
      "base_url":"http://public-repo-1.hortonworks.com/HDP-UTILS-1.1.0.22/repos/ubuntu16",
      "verify_base_url":true
   }
}

hostmap.json

{
  "blueprint":"HACluster",
  "default_password":"admin",
  "host_groups": [
    {
      "name": "services",
      "hosts":
      [
        { "fqdn": "services01" }
      ]
    },
    {
      "name": "datanodes",
      "hosts":
      [
        { "fqdn": "datanode1" },
        { "fqdn": "datanode2" },
        { "fqdn": "datanode3" }
      ]
    },
    {
      "name": "namenode1",
      "hosts":
      [
        { "fqdn": "namenode1" }
      ]
    },
    {
      "name": "namenode2",
      "hosts":
      [
        { "fqdn": "namenode2" }
      ]
    },
    {
      "name": "namenode3",
      "hosts":
      [
        { "fqdn": "namenode3" }
      ]
    }
  ]
}

repo.json

{
   "Repositories":{
      "base_url":"http://public-repo-1.hortonworks.com/ambari/ubuntu16/2.x/updates/2.6.1.5",
      "verify_base_url":true
   }
}

Then I launch this blueprint this way:

$ curl -i -H "X-Requested-By: ambari" -X POST -u admin:admin http://ambariserver:8080/api/v1/blueprints/HACluster -d @cluster_configuration.json
$ curl -i -H "X-Requested-By: ambari" -X PUT -u admin:admin http://ambariserver:8080/api/v1/stacks/HDP/versions/2.6/operating_systems/ubuntu16/repositories/HDP-... -d @repo.json
$ curl -i -H "X-Requested-By: ambari" -X PUT -u admin:admin http://ambariserver:8080/api/v1/stacks/HDP/versions/2.6/operating_systems/ubuntu16/repositories/HDP-... -d @hdputils-repo.json
$ curl -i -H "X-Requested-By: ambari" -X POST -u admin:admin http://ambariserver:8080/api/v1/clusters/HACluster -d @hostmap.json

When everything is launched I get the following image... host are not registered and nothing works... If I delete RESOURCEMANAGER service from one host groups in cluster_configuration.json, everything works properly but "Resource Manager HA" is not deployed as HA...