Deactivating or Removing Existing Versions

About Deactivating and Removing

If you are installing a new version or reinstalling a previous version of SAS In-Database Technologies for Hadoop, you must deactivate or remove other existing parcels or stacks after installing the new one. You can have more than one parcel or stack for a particular product on the cluster, but only one can be active. At a minimum, you deactivate parcels or stacks that you do not want to use. Optionally, you can remove them from the cluster after deactivation.

Cloudera Manager

Example Names

Deactivation and removal of the parcels for SAS In-Database Deployment Package, SAS Quality Knowledge Base, and SAS Data Management Accelerator for Spark each follow the same procedure. The parcel names are SASEP, SASQKB, and SASDMSPARK, respectively. These names are represented in the following procedures by parcel_name. The configuration filenames for the SAS In-Database Deployment Package and SAS Data Management Accelerator for Spark are ep-config.xml and dmp-config.xml, respectively. The index filename for SAS Quality Knowledge Base is default.idx. These filenames are represented in the following procedures by file_name.

Deactivating

To deactivate a parcel using Cloudera Manager, follow these steps:
  1. Log on to Cloudera Manager.
  2. If running, stop any of the parcel_name services:
    1. On the Home page, click the down arrow next to parcel_name service.
    2. Under parcel_name Actions, select Stop, and then click Stop.
  3. Delete the parcel_name service from Cloudera Manager:
    Note: If you are deleting more than one service, delete all services before proceeding to the step of deactivation.
    1. On the Home page, click the down arrow next to parcel_name service.
    2. Click Delete. The parcel_name service no longer appears on the Homethen selectStatus tab.
  4. Deactivate the parcel_name parcel:
    1. Navigate to the Hoststhen selectParcels tab.
    2. For parcel_name, select Actionsthen selectDeactivate. You are prompted either to restart the cluster or close the window.
    3. When prompted, click Close.
      CAUTION:
      Do not restart the cluster.
    4. Click OK to continue the deactivation.

Removing

After deactivating the parcel, follow these steps to remove it:
  1. Remove the parcel_name parcel:
    1. For parcel_name, select Activatethen selectRemove from Hosts.
    2. Click OK to confirm.
  2. For parcel_name, select Distributethen selectDelete.
  3. Click OK to confirm.
    This step deletes the parcel files from the /opt/cloudera/parcel directory.
  4. Manually remove the file_name file:
    1. Log on to HDFS.
      sudo su - root
      su - hdfs | hdfs-userid
      Note: If your cluster is secured with Kerberos, the HDFS user must have a valid Kerberos ticket to access HDFS. This can be done with kinit.
    2. Navigate to the appropriate directory on HDFS.
      • The directory for SASEP and SASDMSPARK is /sas/ep/config/
      • The directory for SASQKB is /sas/qkb/
    3. Delete the file_name file.

Ambari

Example Names

Deactivation and removal of the stacks for SAS In-Database Deployment Package, SAS Quality Knowledge Base, and SAS Data Management Accelerator for Spark each follow the same procedure. The stack names are SASEP, SASQKB, and SASDMSPARK, respectively. These names are represented in the following procedures by stack_service. The configuration filenames for the SAS In-Database Deployment Package and SAS Data Management Accelerator for Spark are ep-config.xml and dmp-config.xml, respectively. The index filename for SAS Quality Knowledge Base is default.idx. These filenames are represented in the following procedures by file_name.

Deactivating

You deactivate a stack by activating another stack.
To deactivate a stack using Ambari, follow these steps:
  1. Log on to the Ambari manager. All deployed versions of the stack_service stack appear in the left pane of the Home page under the allversions text box.
  2. Select the stack_service stack that you want to activate.
  3. Enter the version number of the stack that you want to activate in the activated_version text box on the Configs tab.
  4. Click Save.
  5. Optionally, add a note describing your action, and then click OK.
  6. If you are deactivating more than one stack, finish all deactivation tasks before restarting services.
  7. Click Restart to restart the stack_service after you have deactivated all the stacks.
  8. Click Restart All Affected. The affected services are restarted.
  9. The new stack is activated, leaving the previous stack deactivated.
  10. If you have deactivated additional stacks, select them and restart all affected services. The new stacks are activated, leaving the previous stacks deactivated.

Removing

Note: Root or passwordless sudo access is required to remove the stack.
After deactivating the stack, follow these steps to remove it:
  1. Navigate to the appropriate Admin/bin/stack directory, where stack represents either indatabase, qkb, or dmspark. These directories are on the Linux machine where SAS In-Database Technologies for Hadoop is downloaded and unzipped.
    A delete_stack.sh file is in each stack directory.
  2. Copy the delete_stack.sh file to a temporary directory where the cluster manager server is located. Here is an example using secure copy.
    scp delete_stack.sh user@cluster-manager-host:/mytemp
  3. Use this command to run the delete script.
    ./delete_stack.sh <Ambari-Admin-User-Name>
  4. Enter the Ambari administrator password at the prompt.
    A message appears that offers options for removal.
  5. Enter one of the options:
    • Enter 1 to remove only the file_name file.
    • Enter 2 to remove a specific version of stack_service.
    • Enter 3 to remove all versions of stack_service.
    You are prompted to restart the Ambari server to complete the removal of the SASEP SERVICE.
  6. Enter y to restart the Ambari server. The stack_service no longer appears.