VMware vSphere vSAN 7
During the previous release (vSAN 6.7 U3) vSAN received some love and nice new additions to the product. We were now able to use Cloud Native Storage (CNS) on our vSAN environments to provision persistent storage and to provide visibility into container volumes through the vSphere Client. This could be coupled with vSAN’s powerful storage policy-based management (SPBM) capabilities to further increase operational efficiency and management. Next to this we were given more insights with more dashboards, vROPS capabilities and WSFC on native vSAN VMDKs support.
VMware vSphere vSAN 7 builds further upon these changes made in vSAN 6.7 U3. The new product enhancements can be put together into three groups:
- Full integration with the vSphere Lifecycle Manager (vLCM) to use a cluster-wide desired state image.
- Ability to update hardware firmware (from Dell and HPE at first) through the vSphere UI.
- The new recommendations engine in vLCM will automatically check the HCL and VCG to check for inconsistencies or unsupported configurations.
vSAN Native File Services
This release marks the era in which we can finally use vSAN cluster capacity for file shares.
- We can now use the NFS v4.1 and v3 protocol to serve file shares to virtual machines.
- This new functionality also supports quotas.
- It also works with common vSAN/vSphere features, but SBPM doesn’t work yet.
- It’s also still not possible to use these NFS exports to run virtual machines on.
- It does however support VVOLs.
- And of course, it is also suited to be used with Cloud Native workloads to create persistent-volumes running directly on vSAN.
- This is a cluster level service, which means it has to be enabled on each cluster separately.
Enhanced Cloud Native Storage (CNS) capabilities
- VMware has not stopped developing the Cloud Native Storage capabilities which was released in the previous version of vSAN. They continued the development and are now able to deliver an even richer capability set to the VI-Admins.
- The ability to deliver file-based persistent volumes on vSAN for Kubernetes instances like I said earlier.
- These persistent volumes can also be encrypted and snapshotted.
- These persistent volumes can be resized.
Next to these three key product enhancements VMware also added some other noteworthy additional enhancements to the product.
Improved VM Placement Intelligence with Stretched-Clusters
VMware has added a smart integration of DRS awareness with VMware vSAN Stretched Cluster topologies. In previous versions of vSAN whenever a failed preferred site came back from the dead DRS would instantly uphold the VM placement policies. This meant that once a site came back from the dead, the ISL between these sites would get blown full with I/O’s because of the resync process, but also because of the DRS/vMotion process that would be running simultaneously. With this version VMware tries to reduce the ISL traffic and potential performance impact after a failure by instructing DRS not to uphold the VM placement policies until the resyncs with the other site are done and the virtual machines are compliant again.
The Replace Witness workflow is enhanced
In the previous version it was already pretty easy to replace a vSAN Witness host (which is essentially an .OVA appliance) right from the vSphere Client. There was one flaw with this flow though, which was that it didn’t instruct the vSAN clusters to initiate a repair to regain compliancy. This has changes. Once you change the witness vSAN will get instructed to run an immediate repair to regain compliancy instead of waiting for a timeout. This will essentially minimize interruption that could occur when executing this workflow.
Intelligent Capacity Management in Stretched Cluster Topologies
This new enhancement made to vSAN is useful when you are potentially going to run into capacity issues on one of the sites. Let’s suppose we have two sites in a Stretched Cluster topology. One of the sites has more vSAN capacity in use than the other (because of using flexible per-VMDK storage policies for example). When a site reaches a 100% used capacity, it will impact VM uptime right? Well with this new enhancement vSAN acts on this and essentially redirects active I/O to the other site (Which is not having capacity issues). It does this by marking the component as “Absent”. This will make the VM degraded, but it will increase the VMs uptime because ongoing operations will be able to continue.
More accurate VM Capacity reporting
vSAN finally fixed the inconsistent VM-level capacity storage usage reporting across the vCenter UI and APIs. Previously you could receive different values within different tools (vCenter, vROPS, vSAN etc). This has been fixed and there is now a consistent VM capacity reporting across the product lines.
New vSAN Memory usage Metric
In this release we can now accurately see how much memory the vSAN services are using on an ESXi host. Great new addition if you ask me, we can now see how much memory we are using or are going to be using on ESXi host for vSAN services. This data is available through the vSphere Client and through APIs.
vSphere Replication with vSAN enhancements
vSAN is now fully aware of vSphere Replication data. In previous versions the vSphere Replication objects were unknown to the vSAN environment. With vSAN 7 they are fully visible and identified in the UI. This also means you can view capacity reporting for all of the vSphere Replication objects known to the cluster.
NVMe hotplug support
With this release, vSphere and vSAN will gain hot plug support for NVMe devices. This is a great addition because it will reduce the maintenance windows, minimize host restarts and reduce complexity in servicing vSAN, or vSphere hosts. A side-note for this is that it will only work with a couple of OEM platforms.
Elimination of Eager Zero Thick (EZT) devices for Oracle RAC
Although already “fixed” in vSAN 6.7 P01, I found it noteworthy to mention that VMware has removed the EZT requirement for Oracle RAC VM’s on vSAN. You can use Think Provisioned disks with Oracle RAC without any problems now.
Increased maximums and other changes
- vSAN 7 will now support 32TB physical capacity devices. Which means that with de-duplication and compression you could go over 1PB in logical capacity per vSAN cluster.
- Because of this it could potentially also improve the deduplication ratio within a diskgroup.
- You can now easily override the default gateway used by vSAN hosts. This in return also simplifies deployments for routed or greenfield environments.
- VMware Skyline now has integration with vSphere Health and vSAN Health.
This concludes the changes made to VMware vSAN 7.0 that were noteworthy to report on. Because there is a lot more to come, please continue on to page 3 to find out more on vSphere 7.0 and its new enhancements along the complete product portfolio.
** Please make note that the announced products haven’t been released yet and that features or certain enhancements could potentially be removed or changed before release.