The OpenStack block storage service (Cinder) allows users to attach volumes to their instances. VMware Integrated OpenStack simplifies storage management tasks so that users can be productive much quicker.
One way that we simplify management is by allocating storage to Nova compute clusters using VMware vSphere data stores. The VMware vCenter VMDK driver can utilize Virtual SAN (VSAN), VMware File System (VMFS), and Network File System (NFS) data stores. This significantly reduces the amount of information that the OpenStack administrator needs to specify in the nova, cinder, and glance configuration files. The administrator simply specifies the VMDK driver and the data store names.
VMware Integrated OpenStack provides a further simplification to storage management: wizard-driven storage allocation within the VMware vSphere Web Client! See Figure 1 for an example.
Figure 1: OpenStack Storage Allocation
This allows administrators to automate the necessary changes to the OpenStack configuration files mentioned earlier and removes potential sources of error in the deployment.
Finally, VIO fully supports Storage Policy-Based Management (SPBM). See Figure 2 for a brief overview.
Figure 2: VMware SPBM OpenStack Integration
Storage policies are applied normally to VMware vSphere data stores by administrators. Then, administrators create custom volume types in OpenStack that correspond to the vSphere storage policies. When OpenStack users create their cinder volumes, they specify an extra specification with the desired volume type, and VMware vSphere will create the volume on the appropriate data store according to the policy.
The following video provides a detailed walkthrough of using the OpenStack block storage service.
Stay tuned for the next installment covering the OpenStack Image service! In the meantime, you can learn more on the VMware Product Walkthrough site and on the VMware Integrated OpenStack product page.
Is it possible to use standard OpenStack cinder drivers in this deployment? Didn’t see any mention of that. Will it be possible (for instance) to use a VMAX as a storage system instead of (or in addition to) vSphere datastores?
Hello Anil,
The EMC VMAX can be used with the VMware VMDK driver for Cinder. Just allocate a vSphere datastore on the array and assign it to VIO.
VIO supports the VMDK Cinder driver only. “Why?” you may ask? Well, when we use the VMDK driver, there are a couple benefits:
1. Breadth of hardware support: Any array manufacturer and model that is supported by vSphere is automatically supported by VIO
2. Storage administration is simplified: There is no need to configure per-array drivers if you have multiple array models, which some customers do
3. Storage performance is improved: VMware has invested significant R&D in joint work with lead storage industry vendors to develop VAAI, which offloads time-consuming activities like cloning to the array itself. I am not sure of the status of something similar being implemented in OpenStack array-specific drivers, but this capability is available today in vSphere and VIO.
I hope this helps.
Hi,
I am running VIO 2.0. I am trying to add Nova Storage. I receive this error:
“The operation is allowed only when the deployment is running or in a configuration error state”
I cannot find any documentation on adding more storage in 2.0 but I do in 2.5
http://pubs.vmware.com/integrated-openstack-25/index.jsp?topic=%2Fcom.vmware.openstack.admin.doc%2FGUID-E539CD25-C485-4395-986E-5E3ED9E099D1.html
What is this statement mean? Will my Instances/VMs stay up and running?
Adding a datastore to the Compute node causes the Nova service to restart, which might cause a temporary disruption to the OpenStack services in general.
Thanks
Jeff