As mentioned in my previous vSAN post here, I talked about how important Storage Policies are when it comes to vSAN.
Storage Policy-based Management, or SPBM for short is the secret sauce that binds everything together in vSAN today. With this, you will not need to go through the very manual and inefficient way of setting up storage LUNs.
You manage performance and protection profiles of your workloads at the software layer. What’s important is that we can actually apply these profiles at a granular level- we can apply different policies to different disks of a single VM.
If you ever need to change policies, it is unlike the traditional way where you need to migrate your VMs from one LUN to another. In a software-defined world, everything is pretty much straightforward. All you need is really a few clicks of a button, there is no downtime.
More About SPBM
A policy is made up of several rules. There are a ton of rules relating to performance or protection which you can set.
These are some of the most common ones:
- Failures to Tolerate (FTT)– Like the name says, this rule is all about tolerating n number of failures. You are able to set values of 0 to a maximum of 3 for this rule. Do take note that as the number of FTT you set increases, you are able to tolerate more failures but at the same time, you require more nodes for this.By default, the value will be at 1.
- Object Space Reservation- This rule allows you to set the logical size of the virtual machine disk (vmdk) object. By default, it will be at 0%, where it is thin provisioned. If you increase it to 100%, it means that the vmdk will be thickly provisioned
- Number of Disk Stripes per Object- This rule determines the stripe width across all replicas of a VM.By default the value is 1. The maximum value is 12.Do take note that a value of 1 might equate to better performance, but it also requires you to use more resources. I used the word “might” because in certain situations, increasing the stripe width may backfire and may not necessarily increase performance. More about this later.
There are of course, a lot of other rules in which you can set. For the exhaustive list, you can refer to any of the official VMware blogs or VMware’s official documentations.