IKO Plus: Database Management, Migration, or Day 1 Seeding your Mirrored IrisCluster

Mirror Your Database Across the Galaxy with Seeding

Hello cpf fans! This walkthrough explores using the “seed” capability in IRIS to provision an entire IrisCluster mirror, 4 maps wide, with compute starting from an IRIS.DAT in a galaxy far, far away.

This is incredibly powerful if you have a successful monolithic implementation and want to scale it to the outer rim using Kubernetes and the InterSystems Kubernetes Operator (IKO). Even though my midichlorian count is admittedly low, I’ve seen hardcore Caché hackers shovel around DATs, compact/shrink them, and update ZROUTINES. This same approach can help shrink and secure your containerized workloads too.

---

The Mission

I’ll show you how I built an IrisCluster that pulls an IRIS.DAT file from an arbitrary location via an initContainer and uses it to “seed” a mirrored database throughout its topology.

I used a repository I visit annually: @Guillaume Rongier’s InstallSamples, which features a committed IRIS.DAT.

🌱 This will be the seed for our IrisCluster.

Provisioning the Cluster

First, I provision a quick Kind cluster with 5 worker nodes, install Cilium as the CNI, and deploy the IKO. Then, I label the nodes with Star Wars planet names.

ikoseedcluster.sh

cat <<EOF | kind create cluster --name ikoseed --config=-
kind: Cluster
apiVersion: kind.x-k8s.io/v1alpha4
nodes:
- role: control-plane
- role: worker
- role: worker
- role: worker
- role: worker
- role: worker
networking:
  disableDefaultCNI: true
EOF

kind get kubeconfig --name ikoseed > ikoseed.kubeconfig
cilium install --version v1.18.0 --kubeconfig ikoseed.kubeconfig
cilium status --wait --kubeconfig ikoseed.kubeconfig
helm install iko iris-operator/ --kubeconfig ikoseed.kubeconfig

# Label the planets
planets=(tatooine coruscant hoth naboo endor)
i=0
for node in $(kubectl get nodes --no-headers --kubeconfig ikoseed.kubeconfig | awk '/worker/{print $1}'); do
  planet=${planets[$i]}
  kubectl label node "$node" topology.kubernetes.io/zone="$planet" --overwrite --kubeconfig ikoseed.kubeconfig
  echo "$node → zone=$planet"
  i=$(( (i+1) % ${#planets[@]} ))
done

I now have a happy Kubernetes cluster, nodes labelled as tatooine, coruscant, hoth, naboo, and endor.

---

IrisCluster Topology

Here is the topology declaration, including the critical initContainer and seed configuration.

iriscluster.yaml

apiVersion: intersystems.com/v1alpha1
kind: IrisCluster
metadata:
  name: ikoplus-seed-sweenx12
  namespace: ikoplus
spec:
  imagePullSecrets:
    - name: containers-pull-secret
  licenseKeySecret:
    name: license-key-secret
  volumes:
    - name: airgapdir
      emptyDir: {}
  topology:
    data:
      image: containers.intersystems.com/intersystems/irishealth:2025.1
      mirrorMap: primary,backup,drasync,rwrasync
      mirrored: true
      volumeMounts:
        - name: airgapdir
          mountPath: /airgapdir
      irisDatabases:
        - name: SAMPLES
          directory: /irissys/data/IRIS/mgr/SAMPLES
          mirrored: true
          ecp: true
          seed: /airgapdir/
          logicalOnly: false
      preferredZones:
        - tatooine
        - coruscant
        - hoth
        - naboo
      initContainers:
        - name: init-grongierisc-samples
          image: debian:bookworm-slim
          command:
            - sh
            - -c
            - |
              echo "Retrieving Samples from @grongierisc's repository..."
              apt-get update && apt-get install -y curl
              cd /airgapdir
              curl -L https://media.githubusercontent.com/media/grongierisc/InstallSamples/refs/heads/master/samples/IRIS.DAT --output IRIS.DAT
              chmod -R 777 /airgapdir
              chown -R 51773:51773 /airgapdir
          volumeMounts:
            - name: airgapdir
              mountPath: /airgapdir
          securityContext:
            runAsUser: 0
    compute:
      image: containers.intersystems.com/intersystems/irishealth:2025.1
      replicas: 1
      preferredZones:
        - endor

How Seeding Works

The init-container retrieves the database from GitHub and places it in /airgapdir/IRIS.DAT. In the IrisCluster spec, we tell the operator to use this as a seed:

      irisDatabases:
        - name: SAMPLES
          directory: /irissys/data/IRIS/mgr/SAMPLES
          seed: /airgapdir/

Under the hood, IKO generates a cpf merge with an Actions directive to handle the heavy seeding:

[Actions]
CreateDatabase:Name=SAMPLES,Directory=/irissys/data/IRIS/mgr,Seed=/airgapdir
CreateNamespace:Name=ENSDEMO,Globals=SAMPLES,Routines=SAMPLES,Interop=1

---

Deployment & Attestation

Applying the configuration:

kubectl apply -f iriscluster.yaml -n ikoplus

After a few minutes, the cluster comes to life. I pulled up the Mirroring Monitor to see if our seeded database “SAMPLES” made it to the party.

🎉 It DID!

I also verified the filesystem on the pods to ensure the DAT is in all the right places.

And the ECP client (compute node) is successfully connected:

Testing Replication Across the Galaxy

I wrote a global on the Primary mirror and then looped over the topology to ensure it mirrored correctly.

seedmirrorcheck.sh

#!/bin/bash
export KUBECONFIG=ikoseed.kubeconfig
NAMESPACE=${1:-ikoplus}
CMD='iris session IRIS -U ENSDEMO -B <<EOF
ZWRITE ^MayTheForce
HALT
EOF'

for pod in $(kubectl get pods -n "$NAMESPACE" -o jsonpath='{.items[*].metadata.name}'); do
  echo "---- Pod: $pod ----"
  kubectl -n "$NAMESPACE" exec "$pod" -- /bin/sh -c "$CMD"
done

The results are solid:

• ✅ Primary (Persisted)
• ✅ Backup (Persisted)
• ✅ DRAsync (Persisted)
• ✅ RWRAsync (Persisted)
• ✅ Compute (Accessible via ECP)

🎉 The Force is with us!

---

Related Reading:

• IKO Plus: Operator Works From Home
• IKO Plus: VIP in Kubernetes on IrisClusters
• InterSystems Documentation: CreateDatabase CPF Reference