IBM Support

IBM Cloud Pak for Business Automation 21.0.x Known Limitations

General Page

This web page provides a list of known limitations in Cloud Pak for Business Automation 21.0.x. Workarounds are provided where possible.
The limitations shown on this page are subject to updates between releases. You can also find limitations listed in IBM Documentations:
Current releases
This webpage shows known issues for the following versions. Click the name of the version to go to its section on this webpage.
IBM Cloud Pak for Business Automation 21.0.3 IBM Cloud Pak for Business Automation 21.0.2 IBM Cloud Pak for Business Automation 21.0.1
 
IBM Cloud Pak for Business Automation 21.0.3
The following table shows the relevant versions of the product in the Reported version column. You can find the download document for 21.0.3 at IBM Cloud Pak for Business Automation 21.0.3 Download Document. For fixed defects, see Fix list for Cloud Pak for Business Automation 21.0.3.
 

Installation

Limitation Description
Version
The Business Team Service (BTS) operator fails with an image pull error when you are deploying Cloud Pak for Business Automation capabilities.
Symptom:
The BTS pods show an ImagePullBackOff error:
rpc error: code = Unknown desc = unable to retrieve auth token: 
invalid username/password: unauthorized: 
The login credentials are not valid, or your IBM Cloud account is not active.
 
Cause:
The BTS service account does not use the default image pull secret.
Solution:
1.  Find the BTS service account by running the following command:
oc get sa | grep bts
2.  Add the image pull secret to the BTS service account by running the following command:
oc edit sa <BTS_SA>
3. Paste the following code snippet of the image pull secret at the end of the BTS service account, and then save the change:
imagePullSecrets:
 - name: admin.registrykey
4.  Restart the BTS pod.
21.0.3
The IBM Automation foundation (IAF) insights engine fails with an image pull error when you deploy Business Automation Insights (BAI) by using Cloud Pak for Business Automation Red Hat OpenShift Container Platform FormView.
Symptom:
The IAF insights-engine pods show the following error:
rpc error: code = Unknown desc = unable to retrieve auth token:
invalid username/password: unauthorized:
The login credentials are not valid, or your IBM Cloud account is not active.

 
Cause:
The IAF insights-engine service account does not use the default image pull secret.
Solution:
Update the custom resource (CR) to add the image pull secret under shared_configuration:
shared_configuration:
   image_pull_secrets:
   - admin.registrykey
Wait for the operator to take the updated CR and set the image pull secret for the insights-engine service account.
21.0.3
Cannot open any component user interface (UI).
Symptom:
The following message is displayed when you try to access a route like /cpe/acce:
"3.5.0.0 (xxxxxxxxxx)
Cause:
()
The Zen Watcher pod gets stuck.

Solution:
To resolve the issue:
  1. Find the Zen Watcher pod.
    oc get pod | grep zen-watcher
    zen-watcher-7ff698659c-wwntp 1/1 Running
  2. Delete the Zen Watcher pod to restart it.
    oc delete pod zen-watcher-7ff698659c-wwntp
    pod "zen-watcher-7ff698659c-wwntp" deleted
21.0.3
Installing Cloud Pak for Business Automation fails when IBM Operator Catalog and pinned catalogs exists in the Openshift Cluster.
Symptom:
ibm-common-services-operators in ibm-common-services namespace is in Unknown state and All IBM Automation Foundation operators in Failed state.
Cause:
When you execute cp4a-clusteradmin-setup.sh script it did install the supported pinned catalog sources but when installing cp4ba operator, the install plan pick up ibm-operator-catalog , so that the CommonServices csv is installed with wrong version. After the CommonServices operator is installed, it switch it's own subscription to use opencloud-operators (pinned cs catalog), so at this point both CS operator in CP4BA and ibm-common-services namespace is using opencloud-operators, but the csv is incorrect, which cause the CS installation stuck forever.

Solution:
Delete CS operator in both CP4BA & ibm-common-services namespace, install CS operator in CP4BA namespace by selecting "IBMCS Operators" under “provider type/Source" then search for "IBM Cloud Pak foundational services" in operator hub.
21.0.3

Upgrade

Limitation Description
Version
Kafka is not in ready state in reasonable amount of time after upgrade of IAF.
To resolve this issue:
  1. Replace the crossplane-config configmap in the namespace where Crossplane
    is deployed (ibm-common-services) by this configmap (https://github.com/IBM/ibm-crossplane-operator/blob/master/bundle/manifests/crossplane-config_v1_configmap.yaml").
  2. Restart the Crossplane pod and remove the KafkaClaims. The KafkaClaim CR is regenerated.
Upgrading an air-gapped environment from an interim fix that is not the latest causes the Cloud Pak operator to attempt to upgrade to the latest version.
To resolve this issue:
  1. Uninstall the previous version of the Cloud Pak operator.
  2. Install the 21.0.3 version in the OCP console or by running the cloudctl install-operator command.
21.0.3
After you upgrade the Cloud Pak to 21.0.3, the Zookeeper/Kafka/Entity Operator pods show a CrashLoopBackOff error. Run the following command to see the status of the pods:
oc get pod | grep iaf
To resolve the issue, renew the leaf certificates by following the instructions in Changes to CA certificate and key does not automatically rotate Kafka leaf certificates. 21.0.3
The workflow-authoring-baw-case-init-job fails with the error message "Creation of ICM Plugins failed" when an upgraded custom resource includes the following parameters:
    sc_deployment_patterns: "foundation,workflow"
    sc_optional_components: "baw_authoring"
To resolve the issue:
  1. Get the navigator route by running the following command:
    oc get route -n $cp4a_namespace | grep navigator | awk '{print $2}'
    
    Replace $cp4a_namespace with the target namespace.
    An example route is navigator-upgrade.apps.armored.cp.fyre.ibm.com.
  2. Go into the navigator pod, and then to the /opt/ibm/wlp/usr/servers/defaultServer/configDropins/overrides folder.
  3. Copy the zen_oidc.xml file to new_zen_oidc.xml, and then change the urlPattern parameter value from
    urlPattern="navigator-upgrade.apps.armored.cp.fyre.ibm.com"
    to
    urlPattern="://navigator-upgrade.apps.armored.cp.fyre.ibm.com"
    Where the navigator route is the output of the command in step 1.
  4. Save the new_zen_oidc.xml file.
21.0.3
Fixed in interim fix IF002
Limitation Description
Version
Workflow-workstreams starter pattern does not emit Business Automation Workflow or Case events to Business Automation Insights.
The workflow-workstreams starter pattern does not include an event emitter. If you deploy the pattern together with Business Automation Insights (BAI), no Business Process Manager or Case events are sent to BAI. Therefore, you cannot monitor these events in a Kibana or Business Performance Center (BPC) dashboard. You also cannot try Machine Learning Server, including Intelligent Task Prioritization and Workforce Insights.
 
To try the BAI functionalities, you can use the workflow production pattern. The production pattern for workflow authoring still has these functionalities.
21.0.3
Fixed in interim fix IF007
Workflow-workstreams starter pattern does not include Process Federation Server.
Because the starter pattern does not include Process Federation Server:
  • Case features in Workplace are not supported. You can use the Case Client in Business Automation Navigator to handle case work instead.
  • Full text search, Team Performance, and Process Performance are not supported.
To try Process Federation Server for Business Automation Workflow, you can use the workflow production pattern. The production pattern for workflow authoring still has this functionality.
21.0.3
Publishing might fail when you use the Publish menu on the Project Details page.
When you have unpublished versions of a project and there are no differences between the last two versions, you might encounter a publishing problem:
  •  If the last version has an underscore in its name, publishing works, but a new version is created unnecessarily.
  • If there is no underscore in the name, publishing fails.
If you publish by using the Version menu, and not the Project menu, no problem occurs. Likewise, if the project was edited since the last version, no problem occurs.
21.0.3
You can’t see the names of variables that are already mapped. After you map variables, you can’t see the names of the mapped variables when you relaunch the Next step editor for your action, external service, or automation service. 21.0.3
Limitation Description
Version
The 'External Document' document type is not supported when you add a document to a case instance in the Case Details client-side human interface.
Symptom:
In the Case Details client-side human interface, the Documents section is used to add documents to the case. The selection of 'External Document' document type is not supported.  Only 'Document' selection is supported.
Cause:
The 'External Document' document type is a hidden document type that is not meant to be displayed. 'External Document' and a few other hidden document types are not filtered out of the selection due to limitations in Content Management Interoperability Services API that is used to retrieve the document type information.
Solution:
Do not select 'External Document' document type when you add a document in the Documents tab of the Case Details client-side human interface. Use 'Document' document type instead.
21.0.3
Limitation Description
Version
Data extraction from simple tables is fully supported and some limitations exist for complex tables.
For example when you extract data from the summary section of tables, or if some watermarks, text, or other interfering elements exist in your documents. Complex tables are not fully supported.
For more information, see

IBM Automation Document Processing - What are the supported table structures for data extraction in 21.0.3, and which types provide limited support?

21.0.3
In 21.0.3, the checkbox detection is improved by leveraging the output of Object Detection and adding a detection layer. However, some limitations remain for some types of checkboxes.
For more information, see

IBM Automation Document Processing - What are the limitations for checkbox detection in 21.0.3?

21.0.3
In 21.0.3IF9, the default Deep Learning model is missing. This will stop extraction training  from the designer UI on both new and existing projects. 
Symptom:
The deep-learning pods show the following error:
[Errno 2] No such file or directory: 
'/tmp/deep_learning/RetinaNet_OD/PDB1ONT1/checkpoint'
 
Cause:
The default deep learning model does not exist 
Solution:
Update the custom resource (CR) to add the image under ca_configuration:
ca_configuration: 
  global: 
    image: 
      pull_policy: IfNotPresent
      repository: cp.icr.io/cp/cp4a/iadp
      tag: 21.0.3-IF008
Wait for the operator to take the updated CR and revert your images to 21.0.3IF8
Reminder:
Remove this section in 21.0.3IF10 (June 2022) to get latest fixed image
21.0.3
Deploying the second instance of ADP is not supported in an "All Namespaces" enabled OCP cluster.  Symptom:
When you try to deploy a second instance of ADP in an "All Namespaces" enabled OCP cluster, the CP4BA operator will fail to create the aca_ca_crt certificate. You will see the certificate file is expired error in the CP4BA operator log. 

Cause:
The CP4BA operator has a limitation to create a new aca_ca_crt certificate for the second ADP instance in an "All Namespaces" enabled OCP cluster if the ICP4ACluster custom resource instance has the same name as any existing ADP instance.  

Solution:
You must keep one ADP instance per OCP cluster in an "All Namespaces" enabled environment.  
21.0.3

IBM FileNet Content Manager

Limitation Description
Version
You cannot access the user interface of the Content Management Interoperability Services (CMIS) with identity access management (IAM) authentication. Symptom:
You cannot access to the CMIS UI with IAM authentication.

Cause:
Apache chemistry doesn’t have Zen/IAM JWT Token AuthenticationProvider so it causes limitation/issue with IBM CMIS. Recommended to use GraphQL instead of CMIS.

Solution:
Use the CMIS FNCM route (e.g. https://cmis-<project>.<host>/openfncmis_wlp) instead of the CP4BA Zen route (e.g. https://cpd-<project>.<host>/cmis/openfncmis_wlp) when accessing the CMIS application.  See the icp4adeploy-cp4ba-access-info ConfigMap in the OCP console for the cmis-access-info route details. 
21.0.3
Using an incorrect objectClass in a Lightweight Directory Access Protocol (LDAP) User or Group filter can prevent FileNet Content Manager (FNCM) or Business Automation Workflow (BAW) installation.

When you install FNCM or BAW, Content Platform Engine initialization fails at the step of configure FileNet P8 domain and later init FileNet P8 workflow engine failed. If there are multiple expressions in an LDAP User or Group filter, Identity Access Management (IAM) only considers the first one.  If the first expression doesn't match the objectClass used in your LDAP, then the user or group authorization check fails.

To avoid this issue, verify the objectClass properties that are used in your lc_user_filter and lc_group_filter. List only the object classes that exist in your LDAP.  If you have multiple object classes that are used for Users or Groups, list the most common one first.

21.0.3
 

IBM Automation Decision Services

Limitation Description
Version
Users in a group of collaborators cannot access projects in Automation Decision Services.
 
In Business Automation Studio, users can be added as a group of collaborators to work on a project. The users are supposed to have access to the specified project. However, they cannot access any projects in Automation Decision Services as members of a group.

To work on a project in Automation Decision Services, each user must be added individually as a collaborator.
21.0.3
Limitation Description
Version
Business orchestrations are available only in the development environment.
Business orchestrations can't be deployed to other non-production or production environments.  21.0.3
Long names for model artifacts might cause errors. If you specify a long name for an automation service and its operation, you might receive an error.  21.0.3
New data types are composite only.
 
A data type that you create can be composite type only. You can't create simple data items. 21.0.3
Names of data items can't have spaces or special characters  when the data item is used as input to an automation service.
When you define an automation service as the launch option for a business orchestration, the data items used as inputs for the automation service must have names that don't contain special characters or spaces. 21.0.3
The Document type is not supported in the automation service launch definition.
If you specify Automation service as the launch option for a business orchestration, the data items used as inputs for the automation service can't be of type Document. 21.0.3
Document attributes aren't supported in a composite data item of a launch UI. If a composite data item is used in the launch UI and an attribute of the composite type is Document, the orchestration won't launch.  Instead, use a Document data item. 21.0.3
A business orchestration can't initialize a composite data item. 
An automation service output that is mapped to an attribute of an uninitialized composite data item is not supported. A composite data item must be mapped from an output of an automation service or used in a launch UI or a user task (not as a read-only data item) to initialize the composite data item.  21.0.3
Previewing a business orchestration from IBM Business Automation Studio might show you UIs that aren't updated.
Preview your business orchestration from within the orchestration designer to ensure it reflects the latest authored artifacts. Previewing the orchestration from Business Automation Studio might show a launch UI or user tasks that are not up to date. 21.0.3

IBM Business Automation Navigator

Limitation Description
Version
The IBM Content Navigator Sync Client cannot synchronize server changes.

The administrator needs to set the public URL of the Sync services manually, and the URL must be in a specific format.

  1. Get the CPE repository URL. Go to IBM Content Navigator administrator desktop > Connections > Repositories. The URL is in the format http://\{meta.name}-cpe[-stateless]-svc:9080/wsi/FNCEWS40MTOM/. Get the {meta.name} from the URL.
  2. Go to IBM Content Navigator administrator desktop > Sync Services.
  3. Set the Sync services public URL as follows:
    https:// {meta.name}
    -navigator-svc:9443/sync/notify
21.0.3
In an environment enabled for the IBM Cloud Pak Platform UI (Zen), Business Automation Navigator (BAN) does not support connecting to IBM FileNet Content Manager OnDemand repositories.
Line data and AFP2PDF viewers are not supported when documents are viewed in a Content Manager OnDemand repository from a Content Navigator container in a Zen-enabled environment. 21.0.3
In a CP4BA Business Automation Navigator environment, you add or edit a Word or PowerPoint document by using Office for the web. When you close the add or edit window, the document is not checked in automatically. If you add or edit a Word or PowerPoint document by using Office for the web, check in the document manually.
21.0.3

IBM Content Navigator cannot convert custom fonts into a PDF format

  1. In the CR yaml file, add the font setting gdfontpath for ICN configuration: navigator_configuration.icn_production_setting.gdfontpath: "/opt/ibm/wlp/usr/servers/defaultServer/configDropins/overrides/fonts".
  2. Deploy IBM Content Navigator through the custom resource YAML.
  3. Prepare the custom fonts for the container.
  4. Login to  the OCP Console, switch to your target project, and copy the fonts into the BAN container. For example: 
     oc exec -it $navigatorpod bash
     cp -fr /opt/ibm/java/jre/lib/fonts/* /opt/ibm/wlp/usr/servers/defaultServer/configDropins/overrides/fonts
     exit
    # copy custom fonts to container
     oc cp /tmp/fonts/* $navigatorpod:/opt/ibm/wlp/usr/servers/defaultServer/configDropins/overrides/fonts/
21.0.x
 

IBM Business Automation Insights

Limitation Description
Version
Case (icm) event emitter
In the input field of the Source section of the JSON configuration file, you can configure a connection to only one object store. You cannot configure multiple ceObjectStore parameters.
21.0.3
 IBM Cloud Pak for Business Automation 21.0.2
The following table shows the relevant versions of the product in the Reported version column. You can find the download document for 21.0.2 at IBM Cloud Pak for Business Automation 21.0.2 Download Document. For fixed defects, see Fix list for Cloud Pak for Business Automation 21.0.2.
 

Table of contents

General

Limitation Description Version
Restart of enterprise production deployment fails to elect new primary.
When restarting an enterprise production deployment of Cloud Pak for Business Automation that contains foundation, Operational Decision Manager, Automated Decision Services, Business Automation Insights, and resource registry pods, it’s possible that a new primary (or leader) cannot be elected. As a result, three dba-rr prods are deployed:

 
INFO: 49dbc3708e2fdbeb became candidate at term 7
raft2021/05/25 13:04:25 INFO: 49dbc3708e2fdbeb received MsgVoteResp from 49dbc3708e2fdbeb at term 7
raft2021/05/25 13:04:25 INFO: 49dbc3708e2fdbeb [logterm: 1, index: 3] sent MsgVote request to 18091ce978440543 at term 7
raft2021/05/25 13:04:25 INFO: 49dbc3708e2fdbeb [logterm: 1, index: 3] sent MsgVote request to c349aa75b681dc51 at term 7
raft2021/05/25 13:04:26 INFO: 49dbc3708e2fdbeb is starting a new election at term 7
 

The following error can also occur:
 
raft2021/05/25 13:04:28 INFO: 49dbc3708e2fdbeb became leader at term 9
raft2021/05/25 13:04:28 INFO: raft.node: 49dbc3708e2fdbeb elected leader 49dbc3708e2fdbeb at term 9 
2021-05-25 13:04:28.588349 W etcdserver: failed to reach the peerURL (https://dba2102olm-dba-rr-528928c40f.dba2102olm-dba-rr.dba2102rc2fy.svc:2380) of member c349aa75b681dc51 
(Get https://dba2102olm-dba-rr-528928c40f.dba2102olm-dba-rr.dba2102rc2fy.svc:2380/version: 
dial tcp: lookup dba2102olm-dba-rr-528928c40f.dba2102olm-dba-rr.dba2102rc2fy.svc on 172.30.0.10:53: no such host) 2021-05-25 13:04:28.588377 W etcdserver: cannot get the version of member c349aa75b681dc51 
(Get https://dba2102olm-dba-rr-528928c40f.dba2102olm-dba-rr.dba2102rc2fy.svc:2380/version: dial tcp: lookup dba2102olm-dba-rr-528928c40f.dba2102olm-dba-rr.dba2102rc2fy.svc on 172.30.0.10:53: no such host) 
2021-05-25 13:04:28.597551 E rafthttp: failed to dial c349aa75b681dc51 on stream Message 
 (dial tcp: lookup dba2102olm-dba-rr-528928c40f.dba2102olm-dba-rr.dba2102rc2fy.svc on 172.30.0.10:53: no such host) 
2021-05-25 13:04:29.377445 W rafthttp: health check for peer c349aa75b681dc51 could not connect: dial tcp: lookup dba2102olm-dba-rr-528928c40f.dba2102olm-dba-rr.dba2102rc2fy.svc on 172.30.0.10:53: no such host 


To resolve this issue, delete all the resource registry pods:
$ oc get pod o name | grep "-dba-rr" | xargs oc delete 
pod "icp4adeploy-dba-rr-0001fea50c" deleted 
pod "icp4adeploy-dba-rr-097730e4a6" deleted 
pod "icp4adeploy-dba-rr-80417a3c5f" deleted  

The operator rebuild the resource registry cluster with the latest data, to prevent a timeout. The rebuild process can take at least 1 hour to complete. You can verify the rebuild pods with the following command:
$ oc get pod | grep dba-rr 
icp4adeploy-dba-rr-3449822d45  1/1 Running 0 3m20s 
icp4adeploy-dba-rr-368d340766  1/1 Running 0 3m18s 
icp4adeploy-dba-rr-edac344510  1/1 Running 0 3m19s 
 business automations are missing from the IBM Automation home page.
 
Some projects might not appear in the list of recent business automations on the IBM Automation home page. To see the projects you expect, go to Design > Business automations. 21.0.2
Delay in application availability.
It might take as long as an hour for some of the applications to become available again after you deploy IBM Cloud Pak for Business Automation on Red Hat OpenShift on IBM Cloud Multizone Region with portworx for storage by using the internal Key-Value Database Application interface. 21.0.2
Error when opening existing Capture Applications after upgrading.
After upgrading from 21.0.1.x to 21.0.2.x, you need to upgrade the existing Capture applications created in 21.0.1.x deployment. You also need to update the GraphQL AppResource value inside the Application settings in BA Studio from GRAPHQLAPI_SYSTEM to GRAPHQLAPI_SYSTEM_V1. 21.0.2

Installation

Limitation Description
Version
NVIDIA GPU required for Deep Learning in the Automation Document Processing pattern.
The table of infrastructure requirements for Automation Document Processing (see Identifying the infrastructure requirements) states that for Deep Learning, "it is highly recommended to use worker nodes with NVIDIA GPU for better and faster results." This note should state that NVIDIA GPU is required for Deep Learning in the Automation Document Processing pattern. 21.0.2
Unable to include Business Automation Workflow in deployment.
When you deploy Cloud Pak for Business Automation (CP4BA) by using the Operator Hub Form View user interface, you must manually create a secret with the name ibm-entitlement-key if Business Automation Workflow is selected as one of the CP4BA capabilities to be deployed. See step 5 in Preparing the operator and log file storage. 21.0.2
Step missing from silent cluster installation in Demo deployment.
In the Setting up the cluster in silent mode topic for a Demo deployment, a step is missing between steps 1 and 2. If this step is skipped, the installation will hang after IBM Automation foundation deployment and show this message:
FAIL - prerequisites Deployment failed
The missing step is step 3 from the Setting up the cluster with the admin script topic: "Create the ibm-cp4ba-privileged and ibm-cp4ba-anyuid service accounts (SA), and bind the security context constraints (SCC) to control the actions the SA can take and what it can access."
21.0.2
Business Automation Studio documentation correction.
In the System requirements topic, Table 8. Business Automation Studio default requirements shows the default value for App Engine playback under "Pods are licensed for production/non-production" is "Yes/No." It should instead be "No" because the playback engine is always free. 21.0.2
Debilitating regeneration of keystore and truststore.
In production setup, if you do not provide a keystore and a truststore, they are generated every time the environment is started.
This problem also applies to the server.env file, and the root certificate and key of the Certificate Authority. As a consequence of this regeneration, the environment becomes unusable. To avoid this issue, make sure the following folders are persisted in docker volumes: /shared/custom/tls and /shared/custom/env.
21.0.2
DB2 pod (c-db2ucluster-cp4ba-db2u-0) is not ready.
When deploying Demo deployment, occasionally the DB2 pod (c-db2ucluster-cp4ba-db2u-0) is showing a status of "Running".  However, the DB2 pod itself is not ready (showing "Ready 0/1").  For example:
$ oc get pod
NAME                                             READY   STATUS      RESTARTS   AGE
c-db2ucluster-cp4ba-db2u-0                       0/1     Running     0          48m
There is a simple workaround.  Login to OCP cluster using the command line and execute the following commands to workaround the issue:
 
1.  Execute the command below to get the Statefulset name for DB2.  In this case, the DB2 Statefulset is c-db2ucluster-cp4ba-db2u
$ oc get sts
NAME                       READY   AGE
c-db2ucluster-cp4ba-db2u   0/1     48m
2.  Using the DB2 Statefulset name above as an example, execute the following command:
oc patch sts c-db2ucluster-cp4ba-db2u -p='{"spec":{"template":{"spec":{"containers":[{"name":"db2u","tty":false}]}}}}}'
3.  Confirm that DB2 pod is running with "Ready 1/1" after applying the "oc patch" command in step 2 above.  See example below:
$ oc get pod
NAME                                                              READY   STATUS      RESTARTS   AGE
c-db2ucluster-cp4ba-db2u-0                                        1/1     Running     0          48m
Continue with your deployment and verification of the environment.
21.0.2
ImagePullBackOff happens when install Business Automation Application, Workflow or Workstreams.
When install Business Automation Application or Workflow. The image pull failed with ImagePullBackOff state.
If you meet this error you can workaround that by patching service account manually to include the image pull secret with step below.
First of all. Please confirm what pattern and component you installed and the service account names. You can query all of them by commands:
$oc get sa | grep ae-sa
$oc get sa | grep baw-server
$oc get sa | grep pfs
$oc get sa ibm-mls-default-sa
For example:
$ oc get sa | grep pfs
icp4adeploy-pfs-service-account                     2         38h
Then for all the result. Patch that with image pull secret we want to use as below:
$ oc patch serviceaccount/icp4adeploy-pfs-service-account -- type='merge' -p '{"imagePullSecrets": [{"name": "admin.registrykey"}]}'
serviceaccount/icp4adeploy-pfs-service-account patched
After patched all of the service accounts you get. The image will be pulled on next try.
21.0.2
Warning about PodDisruptionBudgetAtLimit.
 
When you install Process Federation Server as part of the Cloud Pak for Automation, you see a warning: The pod disruption budget is preventing further disruption to pods because it is at the minimum allowed level. This warning message occurs because in 21.0.2, the default number of Process Federation Server replicas is 1 for both demo and enterprise patterns, and 1 is the minimum available value for the PodDisruptionBudget. You can ignore the warning. 21.0.2
Issues installing stand-alone Business Automation Workflow on containers

To install Business Automation Workflow on containers, you must always install the latest interim fix for V21.0.2 that includes JR64401 (IF006 or later), or you will get an error because the Business Automation Workflow operator tries to install the IBM Automation foundation.

If you are deploying stand-alone Business Automation Workflow on containers and the  pfs_configuration.elasticsearch parameter is empty, you must configure the embedded Elasticsearch as part of the preparation steps. See Completing configuration for Elasticsearch.

21.0.2
IBM Business Automation Studio
Limitation Description
Version
Incorrect default number of replica requirements for Workflow Authoring.
In table 8 in the System requirements topic, the number of replicas for BAStudio should be 2. 21.0.2
No branch tab.
IBM Business Automation Studio does not show the Branches tab, which is used to manage branches in decision services. The tab is to be included in an upcoming interim fix. The tab is documented in Configuring projects and Publishing decision services as automation services. 21.0.2
Only support for offline deployment.
In IBM Business Automation Studio, you click Business automation > Workflow >Servers and expect to see the Workflow Process Service runtime server. However, the Workflow Process Service runtime server is not listed. Only offline deployment is supported. 21.0.2
IBM Business Automation Workflow
Limitation Description
Version
Tasks don't open in Process Portal, or task data isn't loaded in Workplace after upgrade.

After you upgrade the operator and Workflow Server or Workflow Authoring from 21.0.1 to 21.0.2, you might find that you cannot open tasks in Process Portal, or your task data isn't loaded in Workplace. To resolve this issue, do the following steps:

  1. Set fixed hostnames for each component. See step 1e in Upgrading IBM Business Automation Workflow Runtime. Set fixed hostnames for Workflow Server or Workflow Authoring, Process Federation Server, and Application Engine.
  2. Apply the updated custom resource YAML file without restarting the operator pod.
  3. Wait for the operator to do the upgrade, and check the results.
21.0.2
Incorrect default number of replica requirements for Workflow Authoring.
In table 9 in the System requirements topic, the number of replicas for Workflow Authoring should be 2. 21.0.2
Federation failure.
Process Federation Server does not correctly federate the Business Automation Workflow system when its PostgreSQL database is specified by using the database.jdbc_url property.
21.0.2
Template creation failure.
When you try to create a template from Business automations > Create > Template, and you select a workflow that includes case features, you receive an error message that says, "Your request couldn't be processed," and indicates that you must check the server log files. In the log file, you find the following error:

"E java.lang.IllegalArgumentException: The repository id: null is not valid!"

If you encounter this error, create a template for the first time by going to Business automations > Workflow > Templates.
21.0.2

IBM FileNet Content Manager

Limitation Description
Version
Recommended maintenance for FileNet P8 domain.
Over time, a FileNet P8 domain might be moved between environments (Virtual Servers) or experience an increase or decrease in the number of application server instances (Server Instances). It is recommended that you maintain a current list of Virtual Servers and Server Instances in the Global Configuration Data for the FileNet P8 domain. Follow this procedure to remove obsolete Virtual Server and Server Instance objects: Remove object entries for obsolete Virtual Server and Server Instances from the FileNet P8 domain. 21.0.2

IBM Business Automation Insights

Limitation Description
Version
Unsupported storage class.
Business Automation Insights does not support the ROKS MZR portworx-db-sc storage class.
Use the portworx-shared-sc storage class instead.
21.0.2
Manual setup of dashboard for Content Event Emitter.
Business Performance Center provides template dashboards to explore Decisions and Workflow sample data. However, template dashboards for Content are not available without manual configuration. For instructions to create a dashboard manually for your Content Event Emitter, see Creating and customizing dashboards. 21.0.2
 

IBM Business Automation Navigator

Limitation Description
Version
No support for ROKS MZR with portworx storage.
Task Manager and Aspera integration with Navigator does not work on Red Hat OpenShift on IBM Cloud Multizone Region with portworx storage. 21.0.2

IBM Workflow Process Service

Limitation Description
Version
Broken link to docker-compose.yml file.
 
In the section 2. Configuring the docker-compose YAML file in the topic Installing Workflow Process service, the URL in the GitHub link icp4a/workflow-process-service-samples (https://github.com/icp4a/workflow-process-service-samples/production/docker-compose.yml) is incorrect. The correct URL is https://github.com/icp4a/workflow-process-service-samples/blob/main/production/docker-compose.yml. 21.0.2
Docker-compose impact on Workflow Process Service install.
When docker-compose is down, it might break the Workflow Process Service install. Configure server.env, keystore and truststore to avoid this problem in production. This problem does not occur with docker-compose start or stop. 21.0.2
Documentation missing instructions for downloading from IBM Passport Advantage

In section 3. Running your environment, in the topic Installing Workflow Process service, in step 2, you can also choose to download the Workflow Process Service package (.tgz file) from IBM Passport Advantage.

1. In Passport Advantage, search for "IBM Cloud Pak for Business Automation 21.0.2 Workflow Process Service Multiplatform English (G01HFEN)" and download "CP4BA_21.0.2-wps.tgz". See IBM Cloud Pak for Business Automation 21.0.2 Download Document.

2. Extract the archive.

3. Load the images by changing directory (cd) to the extracted images folder and run the docker load command:

docker load -i workflow-ps-authoring_21.0.2.tar.gz
docker load -i workflow-ps-server_21.0.2.tar.gz

4. (Optional) Push the images to your docker registry. Log in to your docker registry, and push the docker images into your docker registry using the following commands:

docker login <server>
docker tag workflow-ps-server:21.0.2 <server>/workflow-ps-server:21.0.2
docker tag workflow-ps-authoring:21.0.2 <server>/workflow-ps-authoring:21.0.2
docker push <server>/workflow-ps-server:21.0.2
docker push <server>/workflow-ps-authoring:21.0.2

where <server> is the host of the docker image registry that you want to use to pull the images.

For example myregistry.local:5000 or localhost:8080 for a self-hosted registry.

Update the image url tags in the docker-compose.yml file.

21.0.2
Java Virtual Machine crash from Workflow Process Service.
Setting a specific trace to enabled in the embedded Java Runtime Environment from the Workflow Process Service image might cause Java Virtual Machine (JVM) to crash and display a segmentation error:
 
Type=Segmentation error vmState=0x00000000
J9Generic_Signal_Number=00000018 Signal_Number=0000000b Error_Value=00000000 Signal_Code=00000080
Handler1=00007FF6E4FD8470 Handler2=00007FF6E48AD1F0 InaccessibleAddress=0000000000000000
RDI=00007FF6441C5070 RSI=0000000000000000 RAX=D704122B2A002CB5 RBX=000000000109FF00
RCX=00000000426F8790 RDX=0000000000000000 R8=0000000000000000 R9=0000000000000000
R10=FFFFFFFFFFFFFFFF R11=0000000000000000 R12=00007FF6441C5070 R13=00007FF6E52D1800
R14=D70492216E1C7D3D R15=00007FF65FFFC9D0
RIP=00007FF6E5083D53 GS=0000 FS=0000 RSP=00007FF65FFFC600
EFlags=0000000000010202 CS=0033 RBP=00000000000001B8 ERR=0000000000000000
TRAPNO=000000000000000D OLDMASK=0000000000000000 CR2=0000000000000000
 
This crash can be triggered when a specific trace is enabled, but it might not be limited to trace settings.
To minimize the chance of a JVM crash, add the JVM parameter -Xshareclasses:none when you start the container to disable the shared classes cache. Edit the docker-compose.yml file to add JVM_ARGS='-Xshareclasses:none' under the environment section, for example:
environment:
      - LICENSE=accept
      - EXTERNAL_HOSTNAME=localhost
      - EXTERNAL_HTTPS_PORT=9443
      # PFS properties
      - PFS_ENABLE=false
      - JVM_ARGS='-Xshareclasses:none'
21.0.2
The following tables show the relevant versions of the product in the Reported version columns. The versions include 21.0.1 (download page) and 21.0.1 Interim Fix 001 (21.0.1 IF001 readme).
Most of the issues reported for the 21.0.1 release have been resolved. The fixed ones are shown in the section Issues fixed in interim fix 21.0.1 IF001. For information on fixed 21.0.1 defects, see Fix list for Cloud Pak for Business Automation 21.0.1.
 

Table of contents

General

Limitation Description Version
Limited access for federal agencies.
Federal agencies cannot use certain components (see FIPS 140-2 compliance). 21.0.1

Installation

Limitation Description
Version
NVIDIA GPU required for Deep Learning in the Automation Document Processing pattern.
The table of infrastructure requirements for Automation Document Processing (see Identifying the infrastructure requirements) states that for Deep Learning, "it is highly recommended to use worker nodes with NVIDIA GPU for better and faster results." This note should state that NVIDIA GPU is required for Deep Learning in the Automation Document Processing pattern. All

Failure to install on postgresql.

You get errors when you deploy Operational Decision Manager, User Management System, and Business Automation Navigator by using the IBM Cloud version of postgresql.

Follow these instructions to resolve this issue:

  1. Create a role, for example, by using the following command:
    CREATE ROLE ibmcloudpak PASSWORD 'Cl0udPaks!' CREATEDB CREATEROLE INHERIT LOGIN;
  2. Create a database, and make that role the owner of the database:
    CREATE DATABASE icndb owner=ibmcloudpak ...
  3. Use that role to configure the data source instead of the user that is provided by IBM.
21.0.1 IF001
Unable to reinstall Cloud Pak for Business Automation to the same namespace and project.
 
After deleting an instance of Cloud Pak for Business Automation, you cannot reinstall it to the same namespace and project. You must reinstall it in a new namespace and project. 21.0.1
Limitation Description
Version
Lost access to workplace.
 
A workplace cannot be accessed after a few weeks, requiring the pods to be restarted (see Known limitations and issues). 21.0.1
Higher CPU usage.
 
You encounter higher CPU usage on the Business Automation Workflow and User Management Service pods when you move from an on-premises installation to a container. High CPU overhead for Business Automation Workflow on containers can be observed during authentication. 21.0.1
Limitation Description
Version

Default storage class not allowed.

In the following situations, the Business Automation Insights operator role fails to pick the right storage class from the shared_configuration.storage_configuration section of the custom resource:

  • In the demo, if sc_dynamic_storage_classname is configured while sc_medium_file_storage_classname is set to an empty string:
shared_configuration:
  storage_configuration:
    sc_dynamic_storage_classname: my-class
    sc_medium_file_storage_classname: ''
  • In Enterprise, if sc_medium_file_storage_classname is configured while sc_dynamic_storage_classname is set to an empty string:
shared_configuration:
  storage_configuration:
    sc_dynamic_storage_classname: ''
    sc_medium_file_storage_classname: my-class
To resolve these issues:
  • In demo, keep only sc_dynamic_storage_classname:
shared_configuration:
  storage_configuration:
    sc_dynamic_storage_classname: my-class
  • In Enterprise, keep only the configuration of the fast, medium, and slow storage classes:
shared_configuration:
  storage_configuration:
    sc_fast_file_storage_classname: my-fast-class
    sc_medium_file_storage_classname: my-medium-class
    sc_slow_file_storage_classname: my-slow-class
To make these changes, you must switch to the YAML view in IBM Cloud Pak for Business Automation > Create ICP4ACluster to edit the code.
21.0.1 IF001

IBM Business Automation Navigator

Limitation Description
Version
Missing browser desktop.
 
Under certain conditions, the workplace desktop does not appear in Navigator (see No desktops after a Cloud Pak for Business Automation deployment with OCP). 21.0.1
 
This table shows the limitations resolved in interim fix 21.0.1 IF001.
 

Table of contents

General

Limitation Description Version
Continued testing.
The following areas of Cloud Pak for Business Automation 21.0.1 are still being tested:
  • Support for Red Hat OpenShift Container Platform 4.7
  • High availability and disaster recovery
  • Upgrade from previous versions
Limitations found in this testing are reported on this page.
21.0.1

Installation

Limitation Description Version
Installation recommendations.
The following guidelines are provided for installing Cloud Pak for Business Automation 21.0.1:
  • If you are new to the Red Hat OpenShift platform and containers, install 21.0.1 to take advantage of the new features. 
  • If you have been experimenting with a previous version of Cloud Pak for Business Automation, make a new installation by using 21.0.1. 
  • If you are actively developing a solution with a previous version of Cloud Pak for Business Automation containers, do not upgrade until 21.0.1 IF001 is released shortly.
21.0.1
Deploy in demo as administrator.
For the demo deployments, the clusteradmin script is not managing permissions correctly (see

Installing the capabilities by running the deployment script). You must deploy as a Cloud Pak administrator.

21.0.1
Drivers needed for cluster setup.
The documentation for setting up a cluster by running a script is incorrect. You must add Java Database Connectivity (JDBC) drivers to the operator pod. It is not optional. 21.0.1
Upgrade failure.
 
You cannot upgrade from 20.0.3.x to 21.0.1. Wait until the release of 21.0.1 IF001. 21.0.1
Air gap installation issues.
In Installing an air gap by connecting to a bastion compute device, the installation does not work because IBM Common Services 3.7.1 is no longer available. Also, the installation does not work on Red Hat OpenShift Container Platform 4.7. 21.0.1
Demo not for Red Hat OpenShift Container Platform 4.7.
The demo mode does not work on Red Hat OpenShift Container Platform (OCP) 4.7. The demo works on versions 4.6 of Red Hat OpenShift Container Platform and Red Hat OpenShift Kubernetes Service. 21.0.1
Numbers block installation in Db2.
You cannot install Operational Decision Manager (ODM) on a Db2 server if the server has one of the following numbers in its name: 5432, 50000, 3306, 1433, or 1527. These numbers are normally port numbers for ODM. Change the server name to remove or replace the number. 21.0.1
Missing parameters cause installation failure.
This problem occurs for Operator Lifecycle Manager (OLM) installations in Operator Hub. When installing Cloud Pak for Business Automation 21.0.1 with both Business Automation Insights and Automation Decision Services capabilities, the installation fails if the custom resource does not contain the following parameters:
 
spec:
  ads_configuration:
    decision_runtime:
      enabled: true
Before deploying Cloud Pak for Business Automation, insert the parameters above in the YAML viewer of the deployment editor.
21.0.1
Host name not resolvable by DNS.
In Db2 high availability disaster recovery (HADR) environments, where Db2 is identified by a host name that is not resolvable by DNS, it is necessary to configure the hostAliases setting to map the host name to the IP addresses of the primary and standby Db2 servers.
 
Workaround
  1. In the Operator custom resource file, in the dc_ca_datasource section, provide the following values:
    database_servername: The hostname by which the DB2 primary server identifies itself.
    dc_hadr_standby_servername: The hostname by which the DB2 standby server identifies itself.
    database_IP: The IP address of the DB2 primary server.
    dc_hadr_standby_IP: The IP address of the DB2 standby server.
  2. After deployment completes, add the hostAlias settings to the Advanced Data Preparation (ADP) natural language extractor and deep learning pods by doing the following steps:
    1. Access the Red Hat OpenShift Contatiner Platform (OCP) by using the command line or the administrator console.
    2. Edit the deployment named icp4adeploy-natural-language-extractor, for example, oc edit deploy icp4adeploy-natural-language-extractor.

      Note: The prefix icp4adeploy in the name of the deployment might be different in your environment.

    3. Go to the template.spec section and search for the hostAliases section. If hostAliases does not exist, add the following code:
      hostAliases: 
      - ip: IP_OF_PRIMARY_DB2 
        hostnames: 
        - HOSTNAME_OF_PRIMARY_DB2 
      - ip: IP_OF_STANDBY_DB2 
        hostnames: 
        - HOSTNAME_OF_STANDBY_DB2 
    4. Replace the HOSTNAME_OF_PRIMARY_DB2 and HOSTNAME_OF_STANDBY_DB2 with the respective hostnames that identify the Db2 servers.
    5. Replace the IP_OF_PRIMARY_DB2 and IP_OF_STANDBY_DB2 with the respective IP addresses of the Db2 servers.
  3. If you are using the deep learning component of ADP, repeat the same procedure for the deep learning deployment, for example, oc edit deploy icp4adeploy-deep-learning.
  4. After you save your updates, wait for the natural language and deep learning pods to restart.
21.0.1
Limitation Description Version
Redirect to incorrect login page.
When you log out of a Cloud Pak for Business Automation component and log back in, you are redirected to the Administration Hub. If you try to log into this hub without authorization, you receive HTTP Error 403. After you log out, you must navigate to the application page where you want to sign in. 21.0.1
Misspelled string might cause failure.
A string value is misspelled in the Business Automation Studio database type in the Operator Hub Form View. This error can cause the studio to fail to deploy with PostgreSQL. To resolve the error, change posgresql to postgresql in the YAML view: bastudio_configuration.database.type: postgresql. 21.0.1
Limitation Description Version
Authoring pod restarts.
 
The authoring pod restarts regularly because of a memory leak. 21.0.1
Business Automation Insights needed.
The Process Federation Server component does not work correctly unless you also install Business Automation Insights. 21.0.1
Special characters in oidcClientPassword or Resource Registry reader password can cause issues.
Business Automation Workflow might not work after an enterprise deployment on Red Hat OpenShift on IBM Cloud if there is a special character in the resource registry reader password or the oidcClientPassword parameter in the Process Federation Server admin secret. 21.0.1
Access token validation performance.
High CPU overhead for Business Automation Workflow on containers can be observed during authentication. This issue concerns the WebSphere Liberty application server. For each token validation, the OpenID Connect (OIDC) client opens a new HTTPS connection, which requires secure socket layer (SSL) handshakes instead of reusing an existing connection. These SSL handshakes produce CPU overhead on Business Automation Workflow, other OIDC clients, and the User Management Services (UMS) pod. 21.0.1
Unable to call external automation service.
You cannot call an external Business Automation Workflow automation service from a workflow business automation. 21.0.1
Limitation Description Version
Key for air gap mode.
Automation Document Processing does not install in air gap mode unless you provide a fake admin.registrykey name in the container registry. 21.0.1
Load balance doesn't work for GraphQL and CPE.
Content Platform Engine and GraphQL do not apply pod replication when they are used in Automation Document Processing. 21.0.1
Pods stuck in terminating state.
In an Automation Document Processing environment, a failure in a worker node can cause the pods on that node to be stuck in a terminating state. Manually delete the pods that are stuck in a terminating state. 21.0.1
Deep learning training ends in error state.
Deep learning training ends in an error state if a training document contains a comma in its file name. Although you can upload the file to teach the model, the training step results in an error. Resolve the issue by removing the comma from the file name. 21.0.1
Limitation Description Version
Missing property prevents installation.
Business Automation Insights does not install when shared_configuration.image_pull_secrets is absent from the container registry. 21.0.1
Reset memory property.
The Business Automation Insights enterprise deployment fails when task_manager_memory is set incorrectly. Manually set the property to 1,728 megabytes:
task_manager_memory: 1728mb
21.0.1
Limited event handling.
In large scenarios, Business Automation Insights might not be able to handle many events. By default, Elasticsearch is installed with one shard. To fix this issue, increase the number of shards in Elasticsearch. 21.0.1
Error rendering heatmaps.
Business Performance Center might crash with a blank page when rendering heatmaps due to an issue related to the monitoring source ID. If this happens, refresh the browser and select the monitoring source ALL that corresponds to the same domain. Selecting ALL gives you access to the data of all the monitoring sources of the domain. To avoid losing changes to your chart, ensure that you save it regularly. 21.0.1
Team permission to display charts with monitoring sources.
You can grant a team data permission for monitoring sources that have multiple data items, such as Workflow (BPMM) where the monitoring sources have a process name and a version. The team members can view charts that display these monitoring sources, but they cannot create charts that display these monitoring sources. To grant a team permission to display charts with such monitoring sources, you must add the ALL monitoring source. Notably, for Automation Decision Services (ADS), Business Automation Workflow (BPMN), or Workforce Insights data, you must assign the following permissions, respectively: Decisions (ADS) - ALL, Workflow (BPMN) - ALL, and Workforce Insights - ALL. 21.0.1
Workforce Insights computation failure.
The Workforce Insights computation for processes fails when there are no Business Processing Modeling Notation (BPMN) time series format events available in Business Automation Insights Elasticsearch. This problem might cause Business Automation Workflow to fail to emit BPMN events to Elasticsearch. 21.0.1
Lost application events.
 
Emitted events are lost when you use multiple Kafka partitions. Do not set up Business Automation Insights (BAI) to use multiple Kafka partitions, otherwise BAI fails to capture some events, resulting in data loss. 21.0.1
Unreadable secrets.
 
Db2 and Business Automation Insights cannot always read secrets generated by the Operational Decision Manager roles.  21.0.1
Limitation Description Version
Log in issue for Decision Center.
 
You cannot log into Operational Decision Manager Decision Center when you are using a customized webSecurity.xml file. 21.0.1

IBM Content Collector for SAP

Limitation Description Version
Manually complete ICCSAP installation.
 
You must manually complete the installation of ICCSAP (see Setting up the cluster by running a script). 21.0.1

IBM Business Automation Navigator

Limitation Description Version
Verification failure.
 
Business Automation Navigator fails to install because a database connection cannot be verified. To resolve the problem, use the same login credentials for all the navigator SQL databases. 21.0.1
Navigator needs login server.
Navigator does not work unless you connect it to a Lightweight Directory Access Protocol (LDAP) server. Without an LDAP, Navigator has nothing to authenticate usernames and passwords at login. 21.0.1

[{"Type":"MASTER","Line of Business":{"code":"LOB45","label":"Automation"},"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Product":{"code":"SSBYVB","label":"IBM Cloud Pak for Business Automation"},"ARM Category":[{"code":"a8m0z0000001iUCAAY","label":"User Management Service-\u003EUMS Install\\Upgrade\\Setup"}],"ARM Case Number":"","Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"21.0.1;21.0.2;21.0.3"}]

Document Information

Modified date:
11 August 2022

UID

ibm16426995