If you installed a previous version of Automation Document Processing, you can now move
to a newer version.
Before you begin
These steps assume that you have already downloaded the updated artifacts and upgraded the
operator, per the steps in Upgrading the operator by running a script.
Make a backup of the custom resource YAML that you used to deploy your current version of
Automation Document Processing.
About this task
When you upgrade Automation Document Processing, you must also upgrade the FileNet Content
Manager components that you use with document processing. Read through the Upgrading FileNet Content
Manager topic and this topic before you start the upgrade steps.
Plan your upgrade to accommodate all of the related components.
Your upgrade steps might be different depending on whether you are upgrading an authoring
environment or a runtime environment.
After you complete the upgrade, you must update any deployed document processing applications to
the latest version and re-import them in Business Automation Navigator.
The following configuration parameters must be verified, updated, or added for 21.0.2.
Table 1. Automation Document Processing Docker image changes
in 21.0.2
Configuration parameter |
Comment |
processing |
21.0.2 |
extraction |
21.0.2 |
backend |
21.0.2 |
rabbitmq |
21.0.2 |
redis |
21.0.2 |
natural_language_extractor |
21.0.2 |
deep_learning |
21.0.2 |
viewone |
21.0.2 |
cdra |
21.0.2 |
cpds |
21.0.2 |
cds |
21.0.2 |
gitgateway |
21.0.2 |
mongo |
4.2.14 |
dba-keytool-initcontainer |
21.0.2 |
Procedure
To upgrade Document Processing:
- Update your working copy of the custom resource file to include the relevant values for
your new deployment.
- FileNet Content Manager components:
The information for your FileNet Content Manager component
containers is in Upgrading IBM FileNet Content Manager, including the
new image tag values. If you are not adding functionality that requires an object store
initialization, be sure to comment out the initialization and verification sections for FileNet
Content Manager.
- Document processing components:
Update your image tag versions.
In 21.0.2, updates were
made to ensure a consistent hostname for all components. The result is a change to the default route
hostname in some components. For example, if in the prior deployment an
sc_deployment_hostname_suffix
value of xyz.example.com
was used,
the update would be a change from cpds.xyz.example.com
to
cpds-xyz.example.com
.
If you want to keep the hostname that you defined in a previous release, add the hostname
attribute under the
document_processing.container
section. The following example shows the cpds container:
document_processing:
cpds:
hostname: "cpds.test1.9.x.x.x.x.nip.io"
- Back up your databases:
- Content Platform Engine GCD
- Content Platform Engine object store databases
- Content Analyzer base database
- Content Analyzer tenant databases
- Scale down all of your client containers to 0, including the following pods:
- Business Automation Navigator
- Application Engine
- CPDS
- GraphQL
To scale down a specific
deployment:
kubectl scale --replicas=0 deployment my_deployment_name -n my_project_name
Tip: If you have only the document processing pattern and no other capabilities
deployed, you can scale down all deployments in a project by using the following
command:
kubectl scale --replicas=0 deployment $(kubectl get deployment -n my_project_name | awk '{print $1}' | grep -v NAME) -n my_project_name
- Scale down your Content Platform Engine pod to 0.
- Scale down your spbackend pod to 0.
- Obtain the latest Db2 scripts for the new version from the updated
/ACA/configuration-ha directory in your copy of the
cert-kubernetes repository.
-
Update the base Db2 database:
For AIX or Linux, run
UpgradeBaseDB.sh
.
For Microsoft Windows, run
UpgradeBaseDB.bat
.
- Upgrade the tenant Db2 databases:
For AIX or Linux, run
UpgradeTenantDB.sh
.
For Microsoft Windows, run
UpgradeTenantDB.bat
.
- Optional: Load the default data in the project databases.
For
upgrades to 21.0.1, this step is required.
To use the latest default document types, use the
LoadDefaultData.sh script as described in Creating additional project databases.
Important: Running
LoadDefaultData.sh on existing project databases will delete any customizations
and reset everything back to the default state.
If you are unsure whether a project or tenant database is in use for Document Processing, you can
use the following SQL
statement:
SELECT DBNAME, ONTOLOGY FROM CABASEUSER.TENANTINFO t WHERE t.DBSTATUS ='1'
Where
CABASEUSER
is the schema name for the Base database for Document
Processing.
What to do next
Return to Applying the upgraded
custom resource to complete the upgrade of your components by using the operator.When the
operator upgrade is finished, continue to Completing the IBM Automation Document Processing upgrade.