Cloudbreak blueprint for high-availability Hadoop and Hive

[#1]

This is a basic Ambari blueprint for clusters that implement high availability for HDFS, Yarn and Hive for HDP 2.6. This includes Namenode HA, Resourcemanager HA, Hive Metastore server HA and Hive Server HA. It does not implement a high-availability database for the Hive metastore, though, as that is done within the relational database backed itself. It has been developed and tested using Cloudbreak, but should work with Ambari as well.

Hostgroup Layout

Raw Blueprint JSON

{
  "Blueprints": {
    "blueprint_name": "cb24-hdp26-ha",
    "stack_name": "HDP",
    "stack_version": "2.6"
  },
  "settings": [
    {
      "recovery_settings": []
    },
    {
      "service_settings": [
        {
          "name": "HIVE",
          "credential_store_enabled": "false"
        }
      ]
    },
    {
      "component_settings": []
    }
  ],
  "host_groups": [
    {
      "name": "master_mgmt",
      "components": [
        {
          "name": "METRICS_COLLECTOR"
        },
        {
          "name": "METRICS_GRAFANA"
        },
        {
          "name": "ZOOKEEPER_SERVER"
        },
        {
          "name": "JOURNALNODE"
        },
        {
          "name": "INFRA_SOLR"
        },
        {
          "name": "INFRA_SOLR_CLIENT"
        },
        {
          "name": "METRICS_MONITOR"
        },
        {
          "name": "ZOOKEEPER_CLIENT"
        },
        {
          "name": "HDFS_CLIENT"
        },
        {
          "name": "YARN_CLIENT"
        },
        {
          "name": "OOZIE_CLIENT"
        },
        {
          "name": "MAPREDUCE2_CLIENT"
        },
        {
          "name": "HIVE_CLIENT"
        },
        {
          "name": "TEZ_CLIENT"
        }
      ],
      "cardinality": "1"
    },
    {
      "name": "master_nn1",
      "components": [
        {
          "name": "NAMENODE"
        },
        {
          "name": "ZKFC"
        },
        {
          "name": "RESOURCEMANAGER"
        },
        {
          "name": "METRICS_MONITOR"
        },
        {
          "name": "APP_TIMELINE_SERVER"
        },
        {
          "name": "HIVE_METASTORE"
        },
        {
          "name": "HIVE_SERVER"
        },
        {
          "name": "HCAT"
        },
        {
          "name": "WEBHCAT_SERVER"
        },
        {
          "name": "ZOOKEEPER_SERVER"
        },
        {
          "name": "JOURNALNODE"
        },
        {
          "name": "HIVE_CLIENT"
        },
        {
          "name": "HDFS_CLIENT"
        },
        {
          "name": "YARN_CLIENT"
        },
        {
          "name": "OOZIE_CLIENT"
        },
        {
          "name": "ZOOKEEPER_CLIENT"
        },
        {
          "name": "LIVY2_SERVER"
        },
        {
          "name": "SPARK2_CLIENT"
        },
        {
          "name": "MAPREDUCE2_CLIENT"
        },
        {
          "name": "TEZ_CLIENT"
        }
      ],
      "cardinality": "1"
    },
    {
      "name": "master_nn2",
      "components": [
        {
          "name": "NAMENODE"
        },
        {
          "name": "ZKFC"
        },
        {
          "name": "RESOURCEMANAGER"
        },
        {
          "name": "METRICS_MONITOR"
        },
        {
          "name": "HISTORYSERVER"
        },
        {
          "name": "HIVE_METASTORE"
        },
        {
          "name": "HIVE_SERVER"
        },
        {
          "name": "SLIDER"
        },
        {
          "name": "PIG"
        },
        {
          "name": "OOZIE_SERVER"
        },
        {
          "name": "ZOOKEEPER_SERVER"
        },
        {
          "name": "JOURNALNODE"
        },
        {
          "name": "HIVE_CLIENT"
        },
        {
          "name": "HDFS_CLIENT"
        },
        {
          "name": "YARN_CLIENT"
        },
        {
          "name": "OOZIE_CLIENT"
        },
        {
          "name": "ZOOKEEPER_CLIENT"
        },
        {
          "name": "SPARK2_JOBHISTORYSERVER"
        },
        {
          "name": "SPARK2_CLIENT"
        },
        {
          "name": "MAPREDUCE2_CLIENT"
        },
        {
          "name": "TEZ_CLIENT"
        }
      ],
      "cardinality": "1"
    },
    {
      "name": "datanode",
      "components": [
        {
          "name": "HIVE_CLIENT"
        },
        {
          "name": "TEZ_CLIENT"
        },
        {
          "name": "SPARK2_CLIENT"
        },
        {
          "name": "YARN_CLIENT"
        },
        {
          "name": "OOZIE_CLIENT"
        },
        {
          "name": "DATANODE"
        },
        {
          "name": "METRICS_MONITOR"
        },
        {
          "name": "NODEMANAGER"
        }
      ],
      "cardinality": "3+"
    }
  ],
  "configurations": [
    {
      "core-site": {
        "properties": {
          "fs.trash.interval": "4320",
          "fs.defaultFS": "hdfs://mycluster",
          "hadoop.proxyuser.yarn.hosts": "%HOSTGROUP::master_nn1%,%HOSTGROUP::master_nn2%",
          "hadoop.proxyuser.hive.hosts": "%HOSTGROUP::master_nn1%,%HOSTGROUP::master_nn2%",
          "ha.zookeeper.quorum": "%HOSTGROUP::master_nn1%:2181,%HOSTGROUP::master_nn2%:2181,%HOSTGROUP::master_mgmt%:2181"
        }
      }
    },
    {
      "hdfs-site": {
        "properties": {
          "dfs.namenode.safemode.threshold-pct": "0.99",
          "dfs.client.failover.proxy.provider.mycluster": "org.apache.hadoop.hdfs.server.namenode.ha.ConfiguredFailoverProxyProvider",
          "dfs.ha.automatic-failover.enabled": "true",
          "dfs.ha.fencing.methods": "shell(/bin/true)",
          "dfs.ha.namenodes.mycluster": "nn1,nn2",
          "dfs.namenode.http-address": "%HOSTGROUP::master_nn1%:50070",
          "dfs.namenode.http-address.mycluster.nn1": "%HOSTGROUP::master_nn1%:50070",
          "dfs.namenode.http-address.mycluster.nn2": "%HOSTGROUP::master_nn2%:50070",
          "dfs.namenode.https-address": "%HOSTGROUP::master_nn1%:50470",
          "dfs.namenode.https-address.mycluster.nn1": "%HOSTGROUP::master_nn1%:50470",
          "dfs.namenode.https-address.mycluster.nn2": "%HOSTGROUP::master_nn2%:50470",
          "dfs.namenode.rpc-address.mycluster.nn1": "%HOSTGROUP::master_nn1%:8020",
          "dfs.namenode.rpc-address.mycluster.nn2": "%HOSTGROUP::master_nn2%:8020",
          "dfs.namenode.shared.edits.dir": "qjournal://%HOSTGROUP::master_nn1%:8485;%HOSTGROUP::master_nn2%:8485;%HOSTGROUP::master_mgmt%:8485/mycluster",
          "dfs.nameservices": "mycluster"
        }
      }
    },
    {
      "hive-site": {
        "properties": {
          "hive.metastore.uris": "thrift://%HOSTGROUP::master_nn1%:9083,thrift://%HOSTGROUP::master_nn2%:9083",
          "hive.exec.compress.output": "true",
          "hive.merge.mapfiles": "true",
          "hive.server2.tez.initialize.default.sessions": "true",
          "hive.server2.transport.mode": "http"
        }
      }
    },
    {
      "webhcat-site": {
        "properties_attributes": {},
        "properties": {
          "templeton.hive.properties": "hive.metastore.local=false,hive.metastore.uris=thrift://%HOSTGROUP::master_nn1%:9083\,thrift://%HOSTGROUP::master_nn2%:9083,hive.metastore.sasl.enabled=false"
        }
      }
    },
    {
      "mapred-site": {
        "properties": {
          "mapreduce.job.reduce.slowstart.completedmaps": "0.7",
          "mapreduce.map.output.compress": "true",
          "mapreduce.output.fileoutputformat.compress": "true"
        }
      }
    },
    {
      "yarn-site": {
        "properties": {
          "hadoop.registry.rm.enabled": "true",
          "hadoop.registry.zk.quorum": "%HOSTGROUP::master_nn1%:2181,%HOSTGROUP::master_nn2%:2181,%HOSTGROUP::master_mgmt%:2181",
          "yarn.log.server.url": "http://%HOSTGROUP::master_nn2%:19888/jobhistory/logs",
          "yarn.resourcemanager.address": "%HOSTGROUP::master_nn1%:8050",
          "yarn.resourcemanager.admin.address": "%HOSTGROUP::master_nn1%:8141",
          "yarn.resourcemanager.cluster-id": "yarn-cluster",
          "yarn.resourcemanager.ha.automatic-failover.zk-base-path": "/yarn-leader-election",
          "yarn.resourcemanager.ha.enabled": "true",
          "yarn.resourcemanager.ha.rm-ids": "rm1,rm2",
          "yarn.resourcemanager.hostname": "%HOSTGROUP::master_nn1%",
          "yarn.resourcemanager.hostname.rm1": "%HOSTGROUP::master_nn1%",
          "yarn.resourcemanager.hostname.rm2": "%HOSTGROUP::master_nn2%",
          "yarn.resourcemanager.recovery.enabled": "true",
          "yarn.resourcemanager.resource-tracker.address": "%HOSTGROUP::master_nn1%:8025",
          "yarn.resourcemanager.scheduler.address": "%HOSTGROUP::master_nn1%:8030",
          "yarn.resourcemanager.store.class": "org.apache.hadoop.yarn.server.resourcemanager.recovery.ZKRMStateStore",
          "yarn.resourcemanager.webapp.address": "%HOSTGROUP::master_nn1%:8088",
          "yarn.resourcemanager.webapp.address.rm1": "%HOSTGROUP::master_nn1%:8088",
          "yarn.resourcemanager.webapp.address.rm2": "%HOSTGROUP::master_nn2%:8088",
          "yarn.resourcemanager.webapp.https.address": "%HOSTGROUP::master_nn1%:8090",
          "yarn.resourcemanager.webapp.https.address.rm1": "%HOSTGROUP::master_nn1%:8090",
          "yarn.resourcemanager.webapp.https.address.rm2": "%HOSTGROUP::master_nn2%:8090",
          "yarn.timeline-service.address": "%HOSTGROUP::master_nn1%:10200",
          "yarn.timeline-service.webapp.address": "%HOSTGROUP::master_nn1%:8188",
          "yarn.timeline-service.webapp.https.address": "%HOSTGROUP::master_nn1%:8190"
        }
      }
    }
  ]
}

The blueprint file is also attached: cb24-hdp26-ha.txt

[#2]

Please share hostgroup template as well.

Thanks,

Ramu

[#3]

Hi @Ramu Valleti, I have been using Cloudbreak, which takes care of creating the hosts in the cloud and them mapping them to the hostgroups in the blueprint. I may try this blueprint with Ambari sometime next week. I will comment with my results.

Thanks,

John

[#4]

Can you provide a sample blueprint which has the following

1. HBase

2. Kafka

3. Solr

Thanks

K