MANTA Installation and Troubleshooting

5 min readFeb 17, 2023

Metadata — data about your data — holds necessary information that helps you unlock valuable insights. Insights that will allow you to fully understand your data and get rid of anecdote-driven decisions and processes once and for all. MANTA brings intelligence to metadata management by providing an automated solution that helps you drive productivity, gain trust in your data, and accelerate digital transformation.

MANTA Automated Data Lineage for IBM Cloud Pak for Data is available to clients using IBM’s data fabric solution for data governance and privacy. This integration enables clients to add MANTA’s historical, indirect, and technical data lineage capabilities to Watson Knowledge Catalog in IBM Cloud Pak for Data, helping them perform effective impact and root cause analyses, meet common regulatory compliance standards, and gain deeper insights into data quality issues. In addition, the article will help you install MANTA and resolve installation-related issues.

Assumptions

Deploying MANTA on Cloud Pak for Data version 4.5.x or later running on Red Hat OpenShift cluster
Red Hat OpenShift cluster has access to a high-speed internet connection
User has knowledge and experience managing Red Hat OpenShift cluster

Pre-requites:

IBM Cloud Pak for Data 4.5.x or later version running on Red Hat OpenShift
A user with OpenShift cluster and project admin access
Access to the MANTA license file
Bastion host with 1 vCPU and 4 GB RAM with Linux OS
Download and Install the CPD-Cli utility on the bastion host from this link

Step-by-Step Instructions

Step 1 — Enable MANTA in Watson Knowledge Catalog

Enable advanced metadata import provided through MANTA Automated Data Lineage. For example, run the following command to patch the wkc-cr:

cpd-cli manage update-cr \
--component=wkc \
--cpd_instance_ns=${PROJECT_CPD_INSTANCE} \
--patch="{\"enableMANTA\":True}"

After enabling MANTA Automated Data Lineage, wait for 20–30 min and then check if the manta-artemis, manta-dataflow, and manta-keycloak pods are up and running.

Troubleshooting Tips –

Keycloak failed to start after enabling MANTA.

If the manta-keycloak pod fails to start, follow the steps below to make it healthy.

Go to the openshift web console.
Under the deployment section, search for manta-keycloak
Select the manta-keycloak deployment and go to the environment tab
Add the following env variables and save. DB_ADDR is blank, and you need to keep it blank.

Name          Value
DB_VENDOR     h2
DB_ADDR       <empty>

Go to the workload → Pods → search for the manta-keycloak pod and delete it for the force restart.

Step 2 — Download the Cert tool.

Download the cert-tool from https://ibm.box.com/s/90dlyy33kj9o8pty8krkzddxpgek3vll and copy it to your Linux VM.

Extract the contents by running the following:

tar xvf cert-tool.tar

Change directory (cd) into cert-tool and set the executable flag:

chmod +x *.sh

Copy the license key to the cert-tool folder with the name license.key.
When using existing certificates, run the following commands to extract a copy of the cert and key files.

oc get secret external-tls-secret -o=jsonpath=’{.data.cert\.crt}’ | base64 — decode | tee -a “cert.crt”
oc get secret external-tls-secret -o=jsonpath='{.data.cert\.key}' | base64 - decode | tee -a "cert.key"

When using the existing certificate, run the following command to update the following parameters in the config.properties

Vi config.properties

Common: 
project
route
cert-aliasCommon:
cert_filename
cert_key_filename

./install.sh - skip-create
./update_license.sh

Suppose you have run the install.sh command above, then you may skip the below step.
When creating a new certificate, run the following command to update the following parameters in the config.properties :

Vi config.properties

Common:
project
route
cert-alias
cn
DNS

./install.sh
./update_license.sh

Step 3 — Validation

Login to your Cloud Pak for Data, Open an existing project or create a new one.
Click New Asset and click Metadata Import.
You are all set if you see Discovery and Get Lineage options and can select Get Lineage and click next.
Open https://<Your CPD URL>/manta-admin-gui/ and login using MANTA admin credential
The Username is Admin
Run the following command to get the MANTA Admin password

oc get secret manta-credentials -o jsonpath='{.data.MANTA_PASSWORD} {"\n"}'
echo -n 'Enter the password received in above command' | base64 --decode

- Open https://<you CPD url>/manta-dataflow-server/viewer and login using MANTA admin credentials

Troubleshooting

MANTA is disabled though manta pods are up and running.

Run the following command to get ocp route info. Replace namespace with namespace name where IBM Cloud Pak for Data is installed.

oc get routes -n <namespace>

If you find more than one route, there are chances that product-config in OCP has the old CPD routes used during MANTA installation. Therefore, follow the below steps to update product-config.

oc edit configmap product-configmap

Update ZEN_ROUTE: <with the CPD url you are using to access CPD.”
Edit the secret manta-realm-import-wkc and update all URLs to point to the correct one.
Take the back of the following CRs

oc get mantaflow mantaflow-wkc -o yaml > mantaflow-wkc.yaml
oc get mantaauth mantaauth-wkc -o yaml > mantaauth-wkc.yaml

Delete the CRs

oc delete mantaauth mantaauth-wkc
oc delete mantaflow mantaflow-wkc

Redeploy CRs using the backup

oc apply -f mantaauth-wkc.yaml
oc apply -f mantaflow-wkc.yaml

Re-apply the certificate and license, as this is now a clean install. You can refer to step 2 of this document for this step.

MANTA pods fail to start after deleting the CRs

Check the status of manta-adl-operator-controller-manager by running the following command:

oc get deploy -n ibm-common-services

If the deployment is in a crash state, this could be because of Out of memory errors. Increase the memory limit by running:

oc edit deploy manta-adl-operator-controller-manager \ -n ibm-common-services

Scroll down to the resources section and update the memory limits:

YAML

resources:
  limits:
    cpu: 500m
    memory: 512Mi
  requests:
    cpu: 10m
    memory: 512Mi

MANTA admin gui throws 400 bad request errors

Clear the browser cache and retry or try in private mode.