Configuring via dremio.conf

This topic describes how to configure Dremio, in particular, what properties you need to modify in the dremio.conf file and how to accomplish it.

note

The dremio.conf file uses the HOCON syntax.

This topic covers following:

• Configuring Dremio Services
• Configuring Metadata Storage
• Configuring Distributed Storage
• Configuring Zookeeper
• Configuring High Availability
• Configuring Wire Encryption

Disk Space Recommendations

For cluster deployments, we recommend the query spill volume contain 128 GB per node or 2 * executor RAM, whichever is larger. The Dremio administrator should watch for "out of spill space" messages.

Template for dremio.conf

Template for dremio.conf

paths: {

  # the local path for dremio to store data.
  local: ${DREMIO_HOME}"/data"

  # the distributed path Dremio data including job results, downloads, uploads, etc
  dist: "pdfs://"${paths.local}"/pdfs"

  # location for catalog database (if master node)
  db: ${paths.local}/db,

  spilling: [${paths.local}/spill]

  # storage area for the accelerator cache.
  accelerator: ${paths.dist}/accelerator

  # staging area for json and csv ui downloads
  downloads: ${paths.dist}/downloads

  # stores uploaded data associated with user home directories
  uploads: ${paths.dist}/uploads

  # stores data associated with the job results cache.
  results: ${paths.dist}/results

  # shared scratch space for creation of tables.
  scratch: ${paths.dist}/scratch

}

services: {
  coordinator: {
    enabled: true,

    # Auto-upgrade Dremio at startup if needed
    auto-upgrade: false,

    master: {
      enabled: true,
      # configure an embedded ZooKeeper server on the same node as master
      embedded-zookeeper: {
        enabled: true,
        port: 2181,
        path: ${paths.local}/zk
      }
    },

    web: {
      enabled: true,
      port: 9047,
      ssl: {
        # If SSL for communication path between browsers (or REST clients) and Dremio should be enabled.
        enabled: false,

        # Allow for auto-generated certificates if keyStore option is not set
        # Auto-generated self-signed certificates are considered insecure, and this
        # option should be set to false in production environment
        auto-certificate.enabled: true

        # KeyStore and TrustStore settings default to Java keystore and truststore JVM arguments.
        # If needed to be overridden, then change the below properties

        # KeyStore type
        keyStoreType: ${javax.net.ssl.keyStoreType},

        # Path to KeyStore file
        keyStore: ${javax.net.ssl.keyStore},

        # Password to access the keystore file
        keyStorePassword: ${javax.net.ssl.keyStorePassword},

        # Password to access the key
        keyPassword: ${javax.net.ssl.keyPassword},

        # TrustStore type
        trustStoreType: ${javax.net.ssl.trustStoreType},

        # Path to TrustStore file
        trustStore: ${javax.net.ssl.trustStore},

        # Password to access the truststore file
        trustStorePassword: ${javax.net.ssl.trustStorePassword}
      },
      auth: {
        type: "internal", # Possible values are "internal", "ldap" (includes Active directory)
        # LDAP config file for configuration when auth type is "ldap"
        ldap_config: "ldap.json" # file name provided should be in the classpath.
      }
      tokens: {
        cache: {
          # number of tokens to store locally on this coordinator node
          # (set to 0, if all requests should be made directly to the database of tokens)
          size: 100
          # time (in minutes) after which the token needs to be obtained from the token store
          expiration_minutes: 5
        }
      }
    },

    client-endpoint: {
      port: 31010
    },

    scheduler: {
      threads: 24
    },

    command-pool: {
      enabled: true
      size: 0 # 0 defaults to the machine's number of cores
    }
  },

  executor: {
    enabled: true
  },

  flight: {
      enabled: true,
      port: 32010,
      # Authentication mode for the Arrow FlightServer endpoint.
      # There are two modes:
      #       - legacy.arrow.flight.auth
      #       - arrow.flight.auth2
      # legacy.arrow.flight.auth is backwards compatible and will be deprecated in the future.
      # arrow.flight.auth2 is the new and preferred Arrow Flight authentication method.
      auth.mode: "arrow.flight.auth2"
   }

  fabric: {
    port: 45678,

    memory: {
      reservation: 100M
    }
  },

  conduit: {
    # If set to 0, a port is automatically allocated (typically in ephemeral range). Otherwise, the configured value
    # is used.
    port: 0

    ssl: {
      # If SSL for communication path between Dremio instances should be enabled.
      enabled: ${services.fabric.ssl.enabled},

      # Allow for auto-generated certificates if keyStore option is not set
      # Auto-generated self-signed certificates are considered insecure, and this
      # option should be set to false in production environment
      auto-certificate.enabled: ${services.fabric.ssl.auto-certificate.enabled},

      # KeyStore and TrustStore settings default to Java keystore and truststore JVM arguments.
      # If needed to be overridden, then change the below properties

      # KeyStore type
      keyStoreType: ${services.fabric.ssl.keyStoreType},

      # Path to KeyStore file
      keyStore: ${services.fabric.ssl.keyStore},

      # Password to access the keystore file
      keyStorePassword: ${services.fabric.ssl.keyStorePassword},

      # Password to access the key
      keyPassword: ${services.fabric.ssl.keyPassword},

      # TrustStore type
      trustStoreType: ${services.fabric.ssl.trustStoreType},

      # Path to TrustStore file
      trustStore: ${services.fabric.ssl.trustStore},

      # Password to access the truststore file
      trustStorePassword: ${services.fabric.ssl.trustStorePassword}
    }
  }

  # Set up kerberos credentials in server (applicable for both coordinator and executor)
  kerberos: {
    principal: "",
    keytab.file.path: ""
  }

  web-admin: {
    enabled: true,
    # Port, bound to loopback interface, on which the daemon responds to liveness HTTP requests (0 == auto-allocated)
    port: 0
  }
}

provisioning: {
  yarn: {
    jvmoptions: "",
    # list of jars to be added to the main container classpath
    classpath: []
    app {
      # list of jars to be added to the Dremio application classpath
      classpath: []
      # list of jars to be added to the Dremio application classpath (at the front)
      classpath-prefix: []
    }
    watchdog: {
      # how long to wait for server reply before considering it failed
      poll.timeout: 1000ms
      # interval between two server polls
      poll.interval: 10000ms
      # how many consecutive failed attempts before killing server
      missed.polls.before.kill: 6
      # maximum attempts at killing server
      max.kill.attempts: 10
      # how long before reattempting killing server
      kill.reattempt.interval: 1000ms
    }
  }
}

# the zookeeper quorum for the cluster
zookeeper: "localhost:"${services.coordinator.master.embedded-zookeeper.port}
zk.client.session.timeout: 90000


# These system properties are listed here to allow substitution of system property values for DAC Web SSL properties
# listed in services.web.ssl section. Currently we consider only the system properties listed in this file for
# substitution.
javax.net.ssl {
  keyStoreType: "",
  keyStore: "",
  keyStorePassword: "",
  keyPassword: "",
  trustStoreType: "",
  trustStore:"",
  trustStorePassword: ""
}

registration.publish-host: ""