This is the fifth part of my VCAP7-DTM Design exam series. In part 4 I covered the creation of a physical design for vSphere and Horizon components. This time we take a look at section 4 of the blueprint, the creation of a physical design for horizon storage:
Section 4 – Create a Physical Design for Horizon Storage
Objective 4.1 – Create and Optimize a Physical Design for Horizon Infrastructure Storage
Objective 4.2 – Create and Optimize a Physical Design for View Pool Storage
Objective 4.3 – Create and Optimize a Physical Storage Design for Applications
Objective 4.4 – Create and Optimize a Tiered Physical Horizon Storage Design
Objective 4.5 – Integrate Virtual SAN into a Horizon Design
This article is not a comparison between HCI and traditional storage architecture and if you build hosts by yourself or buy Dell EMC’s VxRail or any other vSAN ReadyNode.
Since it is VMware’s strategy to push vSAN and get away from traditional storage, I only cover vSAN. For my VCDX design I will also move away from traditional storage and use vSAN – it’s also my customer’s strategy. The price for flash storage is decreasing constantly and makes a hybrid vSAN architecture less attractive – at least for our use cases.
In general the storage design of a Horizon implementation is very critical. You have to think about capacity, growth capacity, data/object placement, disaster recovery, kind of SSD disks and so on. But in my opinion, HCI or vSAN makes your life a lot easier and simplifies the storage deployment.
If you fail to correctly size the storage and I/O capacity, your customer’s user experience will suffer or the deployment of new desktops is not possible anymore. So, storage performance and sizing is vital for the satisfactory of your customers and their users!
All-Flash or Hybrid Architecture
The first thing you have to figure out and define is the vSAN platform you are going to deploy – All-Flash or hybrid architecture. A All-Flash vSAN configuration aims at delivering very high IOPS with low latencies. Also in a All-Flash configuration you use two different grades of (flash) disks:
lower capacity and higher endurance device for the capacity tier and
more cost-effective and higher capacity disks for the capacity tier
There is no read cache available in a All-Flash configuration as all data is directly read from the capacity tier. Because you aim for extremely high IOPS, make sure you provide a dedicated 10Gb network for the vSAN traffic.
You can enable the deduplication and compression setting (not available when using a hybrid vSAN) in the vSAN cluster to reduce redundant copies of blocks within the same disk group to one copy and to compress the blocks after they have been deduplicated.
Erasure Coding (RAID 5/6 is only available with All-Flash) provides the same level of redundancy as mirroring, but with a reduced capacity requirement. In general, erasure coding means breaking data into multiple pieces and spread them across multiple devices, while adding parity data in the event data gets corrupted or lost. This is a good and short video about this feature:
If you want to use Horizon together with a vSAN Stretched Cluster, there are some limitations. Please read my article VMware Horizon 7.x and vSAN Stretched Cluster.
When using vSAN without further adjustments, your virtual desktops and infrastructure servers are using the default vSAN storage policy. For infrastructure servers this might be okay, but for our desktops we need to create a new policies. Cormac Hogan has very good material about Horizon and vSAN Storage Policies:
HORIZON VIEW 7 ON VSAN – POLICIES REVISITED
HEADS UP! HORIZON VIEW 6.1 AND AF-VSAN POLICIES
The Number of Failures to Tolerate defines the number of host, disk or network failures a storage object can tolerate. This number of Failures to Tolerate (FTT) has the greatest impact on your capacity in a vSAN cluster. Based on your configured availability requirements for a VM, the settings in the policy can lead to a higher consumption on the vSAN datastore (more copies of your data). For “n” failures tolerated, n+1 copies of the object are created and 2n+1 hosts are required.
Consider to configure FTT = 0 for the OS disk for linked-clone floating pools or if you use full-clone non-persistent desktops. If vSAN should experience a failure, only non-persistent data will be lost.
I hope this information was helpful even we didn’t go to deep. If you need to know more about vSAN, then you’ll find tons of documents and other blogs about this technology.
In part 6 I’ll try to give you more information about the design for a Horizon network.
Hi,
I am now half way reading through your articles and sometimes you link to the next part and sometimes you don’t. May I suggest that you revisit these articles and add a link to each next part?
For the rest: good job so far.
Thanks, Rob! I added the links 🙂