Support for iSCSI in VMware Cloud Foundation 9.0
I recently received a query regarding support for iSCSI in VMware Cloud Foundation 9.0. To be exact, the query was related to iSCSI-backed VMFS volumes, and if VCF 9.0 could be built using these volume types. It took a little bit of digging, and a few questions to our product management team, but I finally got an answer to the question. It is multi-faceted, as there is of course the VCF Management Domain and the VCF Workload Domains. There is also the concept of Principal and Supplemental storage to consider. Finally, there is the concept of “greenfield”, new deployments of VCF and “brownfield”, which is the converging of existing vSphere environments into a VCF deployment.
To clarify, here are the definitions of principal and supplemental storage as they appear in the documentation:
-
Principal storage is used during the creation of a workload domain or a vSphere cluster and is capable of running workloads.
-
Supplemental storage can be added after the creation of a workload domain or a vSphere cluster and can run workloads or be used for data at rest storage (such as virtual machine templates, backup data, and ISO images).
Let’s start with greenfield deployments, where the statements are very clear. There are no greenfield workflows for deploying clusters with iSCSI as principal storage in VCF. This is clearly called out in the documentation under the Storage Models section. However it is possible, as the documentation states, to use iSCSI for supplemental storage in new, greenfield deployments.
The next question which arises is if iSCSI can be used in “brownfield” deployments, i.e., existing vSphere Virtual Infrastructure that is being converged into VCF. Again, the documentation is clear. In the section Converging Existing Virtual Infrastructure to a VCF, it states that customers can converge their current vSphere environment to VCF using “any supported vSphere storage type”. This includes iSCSI. So if you have built your vSphere environment and use VMFS volumes which are presented to your vSphere hosts using iSCSI, you can move to VCF and continue to be supported using iSCSI. You can check out the various supported converging scenarios here.
Good news here for customers who use iSCSI in their current vSphere environments and are looking to move to VCF. Oh, and by the way, the exact same support statements apply to using NVMe-oF (TCP or FC) with VMFS, and to using NFSv4.1. These can be treated in the same way as iSCSI. That is, they cannot be used as Principal Storages in “Greenfield” deployments of VCF. They are supported as “Supplemental Storages” in greenfield, as per the documentation referenced above. However, similar to iSCSI, both NVME-oF volumes and NFSv4.1 volume can be part of vSphere environments which are being converged to VCF. Hope that clears up any ambiguity. Thanks for reading.
