Installation

Pre-requisites

  • Hardware - Refer the sizing guide
  • Install Docker for your platform.
  • Download the Acceldata CLI for your platform.
  • Bash unix shell configured for the user.

Installation guide

  1. Create a home folder for Acceldata and move the binary into the folder.

    mkdir acceldata
    mv ./accelo acceldata/accelo
    chmod +x acceldata/accelo
    • It is recommanded to create this home directory in a path accessible to all the admin users, who will be managing Acceldata setup.
    • This directory will contain all Acceldata configuration and data files.
  2. Initialize Acceldata setup

    Execute:

    ./accelo init
    • You must agree to the Acceldata license agreement for the first time to continue the installation.
    1. Now, depending the user permissions you will be asked to do the following:

      If the user is 'root', output will be:

      Enter a path, where to install Acceldata APM: <ENTER_ACCELDATA_INSTALLATION_DIRECTORY_ABSOLUTE_PATH>
      Environment file created successfully.
      IMPORTANT: Please source it to the current shell by running 'source /etc/profile.d/ad.sh' and rerun this command ..
      • What it means?

        • The environment file for Acceldata has been created in the directory '/etc/profile.d/'.
      • What to do?

        • You must source this file to your current shell by running the command 'source /etc/profile.d/ad.sh'

          Execute:

          source /etc/profile.d/ad.sh

      If the user is not 'root', output will be:

      ERROR: Cannot create environment file at '/etc/profile.d/'. Because: open /etc/profile.d/ad.sh: permission denied.
      IMPORTANT: Please move the file '/home/tad/ad.sh' to the path '/etc/profile.d/' and source it to your current shell by 'source /etc/profile.d/ad.sh' and rerun this command ..
      • What it means?

        • The user running 'accelo' binary doesn't have access to '/etc/profile.d/' directory without sudo.
      • What to do?

        • In your present working directory, a file named 'ad.sh' will be created. Move that file to the directory '/etc/profile.d/' and source it to your current shell by running the command 'source /etc/profile.d/ad.sh'

          Execute:

          sudo mv ./ad.sh /etc/profile.d/
          sudo chmod +x /etc/profile.d/ad.sh
          source /etc/profile.d/ad.sh
    2. Execute the 'init' command again:

      accelo init

      Expected output:

      Enter the AccelData ImageTag: <ENTER_THE_CUSTOMER_IMAGE_TAG>
      ✓ Done, AccelData Init Successful
  3. Check the current status

    Execute:

    accelo info

    Expected output:

    AcceloHome: <ACCELDATA_INSTALL_DIRECTORY>
    AccelData Registry: <ACCELDATA_CONTAINER_REGISTRY>
    AccelData ImageTag: <CUSTOMER_IMAGETAG>
    Number of AccelData stacks found in this node: 0
  4. Login to the Acceldata container registry

    Execute:

    accelo login docker

    Expected output:

    Docker installation found ✓
    Docker login succeeded ✓
  5. Deploy Acceldata core components

    accelo deploy core
    ✓ Done, Configuration file generated
    IMPORTANT: Please modify the file '<ACCELDATA_HOME_DIRECTORY>/config/acceldata.conf' as per the requirements and rerun this command.
  6. Edit the file acceldata.conf in the Acceldata config directory and set the values for your cluster

    # This should be the resource manager URL
    yarn.rm.url = "http://rm.ambari.accelo"
    # This should be the MR Job history URL
    mr.job.url = "http://host1.ambari.accelo:19888"
    # The ATS URL, Leave as is if you dont use ATS
    ats.job.url = "http://host1.ambari.accelo:8188"
    # Metastore URI
    hive.metastore.uris = "thrift://host1.ambari.accelo:9083"
    # WebHDFS
    hdfs.webhdfs = "webhdfs://host1.ambari.accelo:50070"
    # Where can we find spark logs , if you are using the default settings , leave this value as IS
    spark.hdfs.path = "/spark2-history"
  7. If the Cluster has Kerberos enabled then set the following flags to true

    #Enable kerberos on this group
    kerberos.enabled = true
    #If metastore is kerberized
    metastore.kerberized = true
    hive.metastore.kerberos.principal = "hive/_HOST@YOUR_COMPANY.COM"
  8. Execute deploy command again

    accelo deploy core
    • If Kerberos was enabled, the CLI will ask your for a Keytab and Principal, Follow the instructions on the CLI to set up Kerberos configuration.
  9. Navigate to http://localhost:4000, Acceldata is ready for use.

  10. Deploy optional Acceldata addons

    Execute:

    accelo deploy addons

    Expected output:

    > [ ] Alerts
    [ ] Anomaly Detection
    [ ] Capacity Forecasting
    [ ] Director
    [ ] FS Analytics
    [ ] Impala Connector
    [ ] Kafka 0.10.2 Connector
    • Select the addons from the list. Use arrow keys to navigate and spacebar to select and press enter.

LDAP configuration

Acceldata can authenticate users with LDAP. To configure LDAP

  1. Execute:

    accelo ldap configure
  2. If instructed to do so, edit the ldap.conf file and set up the ldap server and other required parameters.

  3. Follow the instructions on screen and complete the configuration steps.

  4. Execute:

    accelo deploy addons
    • Select LDAP from the addons list using spacebar and press enter.
    • If LDAP is already checked in the list, just press enter.
  5. Navigate to http://localhost:4000, and login with your LDAP credentials.

Setting up a LogSearch cluster

  1. Execute:

    accelo configure multinode-logsearch
  2. Follow the instructions on screen

  3. Execute:

    accelo deploy multinode-logsearch

Agent deployment

Node agents

Acceldata uses lightweight node agents for cluster and service monitoring.

  1. Execute:

    accelo configure agents
  2. Follow the instructions on screen

  3. Execute:

    accelo deploy agents
    • This above command will give you the list of available agents to deploy.

    • To deploy node agent, execute:

      accelo deploy agents node-agents
    • Similarly, any other agent can be installed.

Other operations

  1. To stop/remove Acceldata addon services, execute:

    accelo deploy addons
    • Just uncheck the addons from the list using spacebar and press enter.
  2. To restart specific Acceldata services, execute:

    accelo restart <SERVICE_NAME>
    • Instead of a specific service name, 'all' can be used to restart all the services.