With vSphere 7 just around the corner, it is the right time to focus on another foundation for everything VMware does: that is vSAN!
Over the last years, vSAN became essential in the race to hyper-converged infrastructure. Today, we’ll focus on the very last version coming up, vSAN 7, in particular some of the top awesome features!
Native File Services
Despite next-gen workloads are becoming part of our daily life, even containers may still need for persistent volumes. Kubernetes is able to use a CSI driver to call a block device, but it doesn’t make sense without multi-writer capabilities. On the other hand, build and maintain a clustered File System shared across containers sounds like the opposite of every cloud design.
This feature provides the ability to provision and manage NFS v3 & NFS v4.1 file shares alongside vSAN block devices. Therefore, no longer need to maintain separated filer nor any VM providing this File Service features.
Native File Services uses the same SPBM model well known by vSAN administrators. Users create shares, select the appropriate Storage Policy and define the configuration (space efficiency, level of resilience, …). That is said, vSAN File being derived from vSAN, it is a cluster-wide setting, enabled or not.
In brief, vSAN Native File Services is not intended to run VM, it is simply not built for that purpose and not supported anyway. However, you should consider it for Cloud Native workload requiring persistent volumes or traditional workloads that still need a NFS share.
vLCM – vSphere Lifecycle Manager
When it comes to Lifecycle Management, people directly think “update”. Despite this is true, it also means so much more for any of us as Virtualization Engineers.
Until now, VUM gave us the ability to maintain our vSphere environment up-to-date in hypervisor level. Although we all know the importance of drivers and the close relationship they maintain with firmware, their management is still a manual process.
This is where vLCM comes into play.
vLCM is all about providing a framework to hardware vendors. With a centric approach, it allows to apply a level of consistency across the hosts within a cluster. To achieve that, vLCM is using a modern approach comparable to the way Ansible or Puppet are working:
This modern approach consists in:
- Defining the desired result you are looking for,
- Build an image based on these conditions (coming from VMware or hardware vendor),
- Apply it to respective host within the cluster.
Let’s take the example of a host within a cluster which has missed drivers provided by the manufacturer. This is not detected as an issue until they’re some problems at some point. Now, with vLCM, we are able to detect that difference, drift, based on the fact that a host is not a the desired state in the cluster.
Stretched Clusters Topologies
Witness Host replacement
VMware improved the replacement and metadata replacement of failed Witness hostwhen working with Stretched or 2-nodes clusters. By adding a “Replace Witness” button in vCenter and enabling an immediate repair job, vSAN 7 makes sure the Witness host will go back to a consistent state as soon as possible.
Nice improvements have been done regarding capacity handling management. When Stretched Cluster, which is really capacity constraint, has almost reached the cluster full condition, the site that sees this situation will mark the related object as “absent”. By doing so, vSAN will allow the related VM to continue processing frontend IO, the outgoing operation will continue, as opposite to an “out of capacity” normal condition.
The object based architecture of vSAN makes it different from traditional storage, including the way capacity is reported. Indeed, the reporting value of the capacity used for a given VM within the vSAN portion of vCenter vs traditional storage wouldn’t really match out and there’s a good reason for that. In fact, the methods used to provide these capacity informations within vCenter all predates vSAN.
vSAN 7 now includes a fully consistent “Used Capacity” report and because it’s all API driven, softwares like vROps will be able to report more accurately.
vSphere 7 introduces the ability to see information related to vSAN’s footprint directly within vCenter UI. In other words, you can show metrics like “consumed memory“, just for vSAN on specific host. Naturally, like other performance information, vCenter can display it over the course time. This infrastructure monitoring, will helps admins plan for future needs.
vSAN 7 modernizes hyperconverged infrastructure by providing administrators a unified storage control plane for both block and file protocols, and provides significant enhancements that make it a great solution for traditional virtual machines as well as cloud-native applications.
vSphere Lifecycle Manager reduces the complexity of monitoring and maintaining infrastructure by consolidating software, driver and firmware update tools and introducing a desired-state model of implementing a desired image. vSAN 7’s file shares enable admins to rapidly provision a file share via a single workflow; integrated file shares also reduce dependencies on expensive third-party solutions and eases lifecycle management.
This is was just a sneak peak of vSAN 7, other additional improvements provide a wide variety of benefits, which we’ll cover in a future “deep dive” article.