How do I uninstall MDM adaptor suites?

Purpose

Perform this procedure to permanently remove one or more MDM adaptor suites from the NSP.

Note: If you are currently upgrading a Release 22.3 or earlier NSP system, you may need to uninstall an adaptor suite from the older system as an upgrade requirement. The procedure steps include release-specific references to support such upgrade-related activity.

Note: In a DR deployment, you perform the adaptor-suite uninstallation in the primary data center; the NSP file-service-app subsequently synchronizes the primary and standby adaptor configurations.

Note: release-ID in a file path has the following format:

R.r.p-rel.version

where

R.r.p is the NSP release, in the form MAJOR.minor.patch

version is a numeric value

Steps
WARNING 

WARNING

Network Management Degradation

After you remove an MDM adaptor suite, you cannot re-install or reinstantiate the adaptor suite using a restart of any pod in the local NSP cluster, or a DR switchover. Consequently, network management may be severely compromised.

Before you attempt to remove an MDM adaptor suite, you must ensure that there are no dependencies on any adaptor in the adaptor suite.

 

Log in as the root or NSP admin user on the NSP deployer host in the standalone or primary data center.


Open a console window.


Perform one of the following.

  1. If the NSP system is at Release 21.11 or earlier, enter the following:

    cd /opt/nsp/NSP-CN-DEP-release-ID/tools/mdm/bin ↵

  2. Enter the following:

    cd /opt/nsp/NSP-CN-DEP-release-ID/NSP-CN-release-ID/tools/mdm/bin ↵


Enter the following to verify that the RESTCONF token for the NSP cluster is valid:

./adaptor-suite.bash --list ↵

Enter the NSP admin user credentials when prompted.

Messages like the following are displayed if the token is valid.

Using log file: log_file

INFO: timestamp -> Listing MDM adaptors suites ...

INFO: timestamp -> Token requested...

INFO: timestamp -> Token acquired

INFO: timestamp -> Listing adaptor suites...

Installed adaptor suites:

list_of_adaptor_suites

INFO: timestamp -> Done.


It is strongly recommended to list the adaptor suites and check for network management dependencies before any adaptor suite is removed.

To list the adaptor suites on the file system, enter the following:

Note: The --list-adaptors option queries the installed adaptors on the MDM server; by contrast, the --list option displays the adaptor suites in the PostgreSQL database. Consequently, until the next periodic NSP cluster data synchronization, an adaptor that you remove may be temporarily shown in the --list option output.

Note: To filter the command output for a specific vendor, pipe the output to grep; for example, the following lists only adaptor suites whose name includes ‘sros’:

adaptor-suite.bash --list | grep sros ↵

./adaptor-suite.bash --list ↵

Enter the NSP admin user credentials when prompted.


Enter the following to remove one or more adaptor suites:

./adaptor-suite.bash --remove zipfile_1 zipfile_2 . . . zipfile_n

where zipfile_1 to zipfile_n are adaptor-suite zip-file specifications; each must include the absolute path

Enter the NSP admin user credentials when prompted.


Enter the NSP admin user credentials.

The adaptors are removed from each running MDM instance.

Note: It may take 30 minutes or more for adaptors to be removed.


If the NSP system is at Release 22.3 or later, enter the following to display the adaptor suites that the --cleanup option in Step 10 removes:

Note: The --list-cleanup option displays the adaptor suites that are uninstalled from the MDM servers but still available in the PostgreSQL database and file-service-app pod.

./adaptor-suite.bash --list-cleanup ↵

Enter the NSP admin user credentials when prompted.


If the NSP system is a DR deployment, verify that the primary and standby file-service-app pods are communicating.

  1. Log in as the root or NSP admin user on the NSP cluster host in the primary NSP cluster.

  2. Enter the following:

    kubectl exec -n $(kubectl get pods -A | awk '/nsp-file-service-app/ {print $1;exit}') -it nsp-file-service-app-0 -- ls /opt/nsp/containers/nspvolume/fileservice/nokia/nsp/mdm/features/suite/↵

    Messages like the following are displayed, and the MDM adaptor zip files are listed.

    Defaulting container name to nsp-file-service-app

    Use 'kubectl describe pod/nsp-file-service-app-0 -n default' to see all of the containers in this pod.

  3. Log in as the root or NSP admin user on the NSP cluster host in the standby NSP cluster.

  4. Open a console window.

  5. Enter the following:

    kubectl exec -n $(kubectl get pods -A | awk '/nsp-file-service-app/ {print $1;exit}') -it nsp-file-service-app-0 -- ls /opt/nsp/containers/nspvolume/fileservice/nokia/nsp/mdm/features/suite/↵

    Messages like the following are displayed, and the MDM adaptor zip files are listed.

    Defaulting container name to nsp-file-service-app

    Use 'kubectl describe pod/nsp-file-service-app-0 -n default' to see all of the containers in this pod.

  6. Compare the lists of adaptor zip files on the primary and standby file-service-app pods.

    If the lists match, the file-service-app pods are communicating correctly.


10 
WARNING 

WARNING

Network Management Degradation

In a DR NSP deployment, performing this step when the adaptors on the primary and standby clusters do not match may severely degrade network management and require a cluster redeployment.

Before you perform the step in a DR deployment, you must be certain that the list of installed adaptors on the standby MDM servers matches the list of installed adaptors on the primary MDM servers.

Enter the following to delete all references to all uninstalled adaptor suites from the PostgreSQL database and file-service-app:

./adaptor-suite.bash --cleanup ↵

Enter the NSP admin user credentials when prompted.

All references to the uninstalled adaptor suites are removed.


11 

Close the console window.

End of steps