General Discussion

As more and more administrators make the switch from a legacy application model to running those very same applications within lightweight containers backed by Kubernetes clusters, disaster recovery and data protection is top of mind to guarantee business continuity and uninterrupted service delivery to end users.

Hitachi has been continuously making the deployment, configuration, and management of persistent storage on Hitachi Virtual Storage Platform (VSP) easier for administrators with products such as Hitachi Storage Plug-in for Containers and Hitachi Storage Plug-in for Prometheus. As of today Storage Plug-in for Containers provides automated creation and allocation of persistent volumes from the VSP backing virtual worker nodes via iSCSI or bare metal worker nodes from the Fibre Channel protocols, while Storage Plug-in for Prometheus provides performance visualization of those persistent volumes. See our past blogs for more information on Storage Plug-in for Containers and Storage Plug-in for Prometheus.

Global-active device is a feature that implements cross-mirrored storage volumes between two VSP systems accepting read and write I/Os on both sides that are continuously updated. If a disk controller failure occurs at one site, the controller at the other site automatically takes over and accepts read and write I/Os.  If an outage occurs, by design global-active device will block I/O to the respective site where the failure occurred by pushing status Pair Split under Error (PSUE) to the volume and redirect all I/O to the alternate site. Aside from HA, global-active device enables administrators to perform volume data migration non-disruptively, and then discontinue global-active device operations after the migration so you can reuse the volumes at both sites. With these features administrators can easily migrate a running OpenShift cluster as a part of rebalancing operations within their datacenter.

 For more information on Hitachi global-active device visit here.

Hitachi Storage Plug-in for Containers has enabled automated global-active device pair creation on VSP storage systems supporting any Kubernetes cluster using Fibre Channel-SCSI for worker nodes, in addition to automatically creating the backing persistent volume and creating and allocating to the appropriate host groups as a tech preview . After administrators have configured global-active device prerequisites and have installed Storage Plug-in for Containers on their respective cluster, allocating an HA persistent volume claim (PVC) is no different than a standard PVC. The following example shows a YAML file for a storage class utilizing the hspc.csi.hitachi.com provisioner with replication type stretched which defines both primary and secondary storage specifications for ports and pools along with quorum disk information.

Hitachi Storage Plug-in for Containers v3.11 will GA in mid-December 2022 and will bring even more enhancements which include:

• Rancher support with both Rancher Kubernetes Engine 1 (RKE1) or Rancher Kubernetes Engine 2 (RKE2)
• Ubuntu 18.04, 20.04 and 22.04 support
• Adaptive Data Reduction (ADR) support within Kubernetes StorageClass
• Service Processor (SVP) redundancy support
• Multiple Kubernetes cluster support via VSP resource groups
• OCP v4.11 support
• Kubernetes v1.24 distribution support

With these enhancements to Storage Plug-in for Containers and global-active device, customers can now support hybrid deployments of Kubernetes running bare metal worker nodes across multiple datacenters to provide HA. The following architecture shows a stretched RedHat OpenShift Cluster on VMware vSphere running across two data centers via a VMware Metro Storage Cluster (vMSC), which in the case of a disaster can fail pods over from datacenter 1 to datacenter 2 for bare metal workers utilizing Fibre Channel in conjunction with Storage Plug-in for Containers. In the following scenario, the third master node is on another cluster as it acts as the arbiter in case of site loss to avoid a split-brain .

As another ROI enhancement, Hitachi has enabled global-active device users to also have the capability of utilizing a cloud provider as quorum device host, which allows customers to not require a third site for quorum and an additional VSP which provides further costs savings. Alternatively, a disk in an iSCSI-attached server can be used as a quorum disk if the server is supported by Universal Volume Manager (UVM). For more information about the Amazon Web Service (AWS) quorum read here.

We look forward to expanding additional capabilities for the Hitachi Virtual Storage Platform and want to thank our customers and partners. For more information on the Cisco and Hitachi Adaptive Solutions program see www.hitachivantara.com/cisco.

Best,

Arvin