End-to-end mirroring of IBM Fusion and its services

If you are planing a disconnected or offline installation, then you must mirror images to your enterprise registry. You can use this procedure to complete an integrated mirroring of IBM Fusion and its services Backup & Restore, IBM Data Cataloging, and Content-Aware Storage. Otherwise, you can mirror them individually per your requirement.

Before you begin

Before you mirror, go through the prerequisites. See Mirroring prerequisites.
Configure Common Environment Variables:
Important: Ensure you review the offline upgrade considerations before proceeding with upgrade mirroring.
Define the following environment variables for your target container registry. These values will be used in each of the subsequent steps to mirror all the related components.
```
export LOCAL_ISF_REGISTRY="<Your container registry host>:<port>"
export LOCAL_ISF_REPOSITORY="<Your image path>"
export TARGET_PATH="$LOCAL_ISF_REGISTRY/$LOCAL_ISF_REPOSITORY"
```
LOCAL_ISF_REGISTRY is your target container registry.
LOCAL_ISF_REPOSITORY is the image path in which you want to mirror the images. You can choose your own repository paths. For example, hci-images/isf or hci-images. See the following sample values:
```
export LOCAL_ISF_REGISTRY="registryhost.com:443"
export LOCAL_ISF_REPOSITORY="mirror-fusion-services-images"
export TARGET_PATH="$LOCAL_ISF_REGISTRY/$LOCAL_ISF_REPOSITORY"
```

About this task

Points to note about this task:

Run all commands as root user.
In the commands, replace <your enterprise registry> with your enterprise registry.
High level tasks to complete this procedure.
1. Install OC CLI and ibm-pak plugin. For steps to install, see Mirroring prerequisites.
2. Set common variables used by all subsequent mirroring steps, such as the target registry and target registry location.
3. Mirror IBM Fusion and related services (Backup & Restore, IBM Data Cataloging, and Content-Aware Storage). For steps, see Procedure.
  1. Set environment variables necessary for mirroring IBM Fusion and related services
  2. Run oc ibm-pak steps to generate the files need to mirror
  3. oc-mirror on the generated files
For the individual procedures, see the following topics:
- Mirror images for Red Hat® operator. See Mirroring Red Hat operator images to enterprise registry.
- IBM Fusion HCI images. See Mirroring IBM Fusion images.
- Mirror Global Data Platform images, see Mirroring IBM Storage Scale images.
- Mirror Fusion Data Foundation images, see Fusion Data Foundation images.
- Backup & Restore images. See Mirroring Backup & Restore images.
- IBM Data Cataloging images. See Mirroring IBM Data Cataloging images.
- CAS images. See Mirroring Content-Aware Storage (CAS) images.

Procedure

Mirror IBM Fusion and related Backup & Restore, IBM Data Cataloging, and CAS services.

Define the following environment variables related to mirroring IBM Fusion and related services.
```
export CASE_NAME=ibm-spectrum-fusion-sds
export CASE_VERSION=2.10.1
```
By default, Backup & Restore, IBM Data Cataloging, and CAS are mirrored together with IBM Fusion in a single set of commands, so you do not have to mirror them separately.
Important: The redhat-oadp-operator and amq-streams operator packages are required only for Backup & Restore and IBM Data Cataloging services.
Ensure that the redhat-oadp-operator and amq-streams operator packages are present in your cluster.

Note: If you have not mirrored redhat-oadp-operator and amq-streams from the Red Hat packages previously, then follow the steps that are provided in the Mirroring Red Hat operator images to enterprise registry.
Note:
- Run the following command to get list of mirrored Red Hat packages available on your cluster.
```
oc get packagemanifests | grep -i "Red Hat Operators"
```
- If you do not get redhat-oadp-operator and amq-streams operator packages from the previous step, then you must follow the step Mirroring Red Hat operator images to enterprise registry.
- Ensure that you also add existing packages along with new one in ImageSetConfiguration file. Otherwise, old packages can be lost from the Red Hat operator index image.
Configure the ibm-pak plugin to use the oc mirror command:
```
oc ibm-pak config mirror-tools -e oc-mirror
```
Use the ibm-pak get command to download the mirroring metadata from IBM’s public CloudPak repository:
```
oc ibm-pak get --version "${CASE_VERSION}" "${CASE_NAME}"
```

Run the ibm-pak generate command to generate the

oc
mirror

configuration files specific to your environment:

oc ibm-pak generate mirror-manifests --version "${CASE_VERSION}" "${CASE_NAME}" "${TARGET_PATH}"

Example output when the generate command completes successfully:

...

- To mirror the non curated catalog:

  oc mirror --config /root/.ibm-pak/data/mirror/${CASE_NAME}/${CASE_VERSION}/image-set-config.yaml docker://${TARGET_PATH}

Note: Use --dest-tls-verify=false parameter when you mirroring images to quay repository.
Run the oc mirror command for the non curated catalog provided in the output of the previous generated command.
Example:
```
 oc mirror --config /root/.ibm-pak/data/mirror/${CASE_NAME}/${CASE_VERSION}/image-set-config.yaml docker://${TARGET_PATH}
```
Example output on successful mirroring:
```
Writing image mapping to oc-mirror-workspace/results-1747667151/mapping.txt
Writing CatalogSource manifests to oc-mirror-workspace/results-1747667151
Writing ICSP manifests to oc-mirror-workspace/results-1747667151
```
In case of any failure, refer Troubleshooting and known issues in offline mirroring to get the resolution steps.
Go to the directory that contains the image-set-config.yaml file that is referenced in the previous oc mirror command. This directory contains several files that were automatically generated by the ibm-pak tool and apply the generated image-digest-mirror-set.yamland catalog-sources-linux-amd64.yaml files to your cluster:
```
cd /root/.ibm-pak/data/mirror/${CASE_NAME}/${CASE_VERSION}/
oc apply -f image-digest-mirror-set.yaml
oc apply -f catalog-sources-linux-amd64.yaml
```
The generate command adds additional catalog sources to the generated catalog-sources.yaml file that is not used. To avoid confusion, it is best to delete these additional catalog sources:
```
oc delete catalogsource -n openshift-marketplace ibm-db2uoperator-catalog ibm-fusion-bnr-catalog ibm-spectrum-discover-catalog
```
If you see Error from server (NotFound) messages, this indicates that the catalogs do not exist and the message can be ignored.
Important: The process also creates a CAS catalogsource as well, and you must not delete it. The delete command also does not clean the CAS catalogsource.