Persistent volumes

The Orient Me, Elasticsearch, and Customizer components for IBM Connections require you to create persistent volumes.

Kubernetes uses persistent volumes (PVs) to store data that needs to be retained regardless of the pod lifecycle. There are many variants and configurations on how to provision persistent volumes; see the Kubernetes documentation to pick the best way for you: https://kubernetes.io/docs/concepts/storage/persistent-volumes.

The Component Pack for Connections requires you to create PVs for different services as described in the following list:

Note: Before creating persistent volumes, ensure that the Prerequisites for Component Pack are completed, including creating the namespace, as the persistent volume claims need to be bound to this namespace.

Mongodb

Mongodb is included in the Infrastructure Helm chart. This helm chart is required for both Customizer and Orient Me. Mongodb supports using PVs, or using StorageClass to provision PVs for you.

If you use PVs for Mongodb, the following PVC names are required:

mongo-persistent-storage-mongo-0
mongo-persistent-storage-mongo-1
mongo-persistent-storage-mongo-2

If you use StorageClass to provision the PVs, append the following statement to the Helm install command when installing the infrastructure chart:

--set mongodb.volumeClaimTemplates.storageClass=storageClassName

Solr

Solr is only included in the Orient Me Helm chart. Solr supports using PVs, or using StorageClass to provision PVs for you.

If you use PVs for Solr, the following PVC names are required:

solr-data-pvc-solr-0
solr-data-pvc-solr-1
solr-data-pvc-solr-2

If you use StorageClass to provision the PVs, append the following statement to the Helm install command when installing the infrastructure chart:

--set solr-basic.volumeClaimTemplates.storageClass=storageClassName

Zookeeper

Zookeeper is only included in the Orient Me Helm chart. Zookeeper supports using PVs, or using StorageClass to provision PVs for you.

If you use PVs for Zookeeper, the following PVC names are required:

zookeeper-data-pvc-zookeeper-0
zookeeper-data-pvc-zookeeper-1
zookeeper-data-pvc-zookeeper-2

If you use StorageClass to provision the PVs, append the following statement to the Helm install command when installing the infrastructure chart:

--set zookeeper.volumeClaimTemplates.storageClass=storageClassName

Elasticsearch

Elasticsearch supports using PVs, or using StorageClass to provision PVs for you.

If you use PVs for Elasticsearch, the following PVC names must be used at the claimRef.name statement:

es-pvc-backup
es-pvc-es-data-0
es-pvc-es-data-1
es-pvc-es-data-2

If you use StorageClass to provision the PVs, append the following statement to the Helm install command when installing the Elasticsearch chart:

--set common.storageClassName=storageClassName

Customizer

The mw-proxy component used by Customizer requires a PV to store customizations in. The following PVC name is required:

customizernfsclaim

Because the mw-proxy component has three pods connecting to one volume, the method used to configure the persistent volume for mw-proxy must support "ReadOnlyMany" (the volume can be mounted read-only by many nodes). Refer to the "ReadOnlyMany" column in Table 1 for the mw-proxy (Customizer) PV support list:

Table 1. Full list of Volume Plugin PV access modes
Volume Plug-in	ReadWriteOnce	ReadOnlyMany	ReadWriteMany
AWSElasticBlockStore
AzureFile
AzureDisk
CephFS
Cinder
FC
FlexVolume
Flocker
GCEPersistentDisk
Glusterfs
HostPath
iSCSI
Quobyte
NFS
RBD
VsphereVolume			Supported when pods are co-located
PortworxVolume
ScaleIO
StorageOS

The Component Pack package includes an optional Helm chart that will set up the PVs using NFS. For instructions on using the provided Helm chart, see the next topic, Setting up persistent volumes with NFS.