A datastore cluster is a grouping of independent datastores designed to function as a single entity. Tasks like virtual machine placement, storage performance optimization, and resource allocation policies can be applied to the cluster and will utilize all datastores in the cluster. When you add a datastore to a datastore cluster object, the datastores resources will be aggregated into the cluster, which will enable you to apply the cluster benefits to the newly added datastore.
Things to Know:
- VMFS and NFS cannot be part of the same datastore cluster.
- Similar disk types should be used inside a datastore cluster
To create the Datastore cluster ->from Web client -> Home screen-> Click “Storage”
Right Click on the datacenter and Choose “New Datastore Cluster”
Enter the datastore cluster name and check if the “Turn on Storage DRS” option is enabled (Its enabled by default) -> Click “Next”
Choose your Storage DRS Automation level (Everything is pretty well descriptive on the web client as shown in the screen shot) ->Click “Next”
As my recommendation. Do not select fully automated level at the start. Choose the default “No Automation level” and leave it for some days. Check the SDRS behavior. If you will satisfied then Choose the fully automated option.
I/O metric is enabled by default and it will trigger in the case of any Storage DRS Threshold condition met. Below showed threshold are by-default and SRDS will occur in either case, If datastore is 80% full or I/O Latency of the datastore in greater than 15 m/s. then VMs will be migrated from the current datastore into the other datastore in the Storage Cluster. If I/O metric was disabled then the initial placement for VMs’ files based on IOPS and IOPS load balancing will not be supported means it will make placement on the base of space utilization.
Default VM Affinity (checked by default, when the VM Migrated between the different Datastore in the Datastore cluster, if it has multiple disk. all disk will be keep together on the same datastore where the VM will move)
Along with above condition it will check I/O imbalance every “8 Hours” and give you recommendation after if source and destination space has 5% space different. Means if the setting was 5% and the source was at 80% utilization, the target would have to have 75% or less capacity utilization before action would be taken.
I/O imbalance have different threshold level, default is fine and tune as per your need. More aggressive mean. Slightly imbalance migration will cause the SDRS to occur.
Choose Host or Cluster -> Click “Next”
Leave the default option “show datastore connected to all hosts”. Unless you have specific use case. Select at least two datastore and make sure it should have same type and disks of the datastore have same attribute (speed, type etc) -> Click “Finish”
Verify your created datastore cluster as shown below.