To model and size storage infrastructure when migrating from VMware Infrastructure, it is essential to understand how to analyze the underlying storage metrics provided by VMware.
There are two key storage containers in VMware and two views in Lanamark One that facilitate analysis of these containers:
- Virtual Disks (connected to Virtual Machines)
The following diagram illustrates the relationship between VMware virtual machines, virtual disks and datastores at different abstraction levels within VMware:
Figure 1. Overall relationship between datastores and underlying Virtual machines per VMware documentation https://pubs.vmware.com/vsphere-50/index.jsp?topic=%2Fcom.vmware.wssdk.pg.doc_50%2FPG_Ch8_Storage.10.3.html
VMware Virtual Datastores View
This view is from the perspective of the storage infrastructure that stores data for the VMware environment and provides an overview of the repository for virtual machine files including log files, scripts, configuration files, virtual disks and other data required for virtual machines to operate properly.
This view takes into account storage optimization such as de-duplication and compression provided by the underlying storage infrastructure.
The following datastore statistics are displayed in the Assessment view of each engagement with VMware Infrastructure:
Figure 2. Virtual Datastore storage statistics available in the Lanamark One Assessment view.
Upon drilling down, the following detailed statistics per datastore are available:
Figure 3. Drilled down view of the datastores.
This detailed view is not VM-aware and only disk usage for each datastore is presented.
On these different views, the following metrics are exposed:
- Total Capacity: Total formatted size of all volumes, and represents the sum of capacity for all datastores in the detail view; it shows how much data could be potentially stored on the datastore.
- Provisioned Capacity: amount of space used across all datastores, including configuration and swap files. Additionally, this is the sum of provisioned capacity in the detail view.This represents the amount of space required for the VMs to function.
- Swap files and snapshots: Swap files and snapshots are required for properly running VMs but are created post-startup. In other words, they are not part of the original capacity of the VM but in addition to it. In some instances, Provisioned Capacity may be more than double the size of total capacity depending on the size of snapshots and required swap files.
- Used Capacity: Total capacity that is consumed across all datastores, and represents the sum of Used in the detail view. Used capacity can be confusing in some cases if remote storage solutions are in use.
Used Capacity represents how much space is physically being consumed on the storage infrastructure, rather than the amount of data stored from the perspective of the guest OS.
Overprovisioning a server
This means running more VMs on a given infrastructure than what can be supported (assuming eventual full allocation). If there is a clear understanding of what the Used Capacity of the virtual machine will be, the storage can be loaded with more VMs than normal.
For example, leveraging storage infrastructure solutions with data de-duplication or compression may report Used Capacity significantly less than the capacity consumed by data stored.
This creates interesting scenarios when it comes to migration planning:
- If migrating to the same storage array with equal or greater storage capacity, datastore Used Capacity will adequately show how much storage is needed.
- If migrating to a different storage array or hyper-converged infrastructure solution, it is important to account for features and storage optimization of the new infrastructure.
- If migrating to the cloud, datastore Used Capacity will not show how much data will need to be stored. In this case, the Used Capacity metric under the VMware Virtual Machines is more suitable for sizing storage requirements.
When is the VMware Virtual Datastore View relevant?
In order to understand the ability of the underlying virtual infrastructure to support existing virtual machines, the datastore view is the most relevant. This does not look into the VM files themselves and thus would not give a clear view of virtual machine utilization and by extension is not an appropriate metric for virtual machine health.
VMware Virtual Machines Storage View
This view provides an overview of storage currently available to all virtual machines and is the sum of total capacity across all virtual disks. It is one level above the virtual datastore infrastructure where virtual machines can only see files that they need to run.
Figure 4. This is the view at which the Virtual Machine storage view operates.
Virtual machine storage metric is composed of whatever is specified in the VM configuration. The datastore information is only relevant for showing the location of files required by VMs.
The following Virtual Machine Storage statistics are available in the Assessment view for VMware Infrastructure:
Figure 5. VMware Virtual Machine storage statistics available in the Lanamark One Assessment View
Figure 6. Drilling down on the Virtual Machine Storage statistics will show results on a per machine basis. Note that there is no mention of underlying storage hardware in this view.
The key fields of interest in the summary view are:
- Total Capacity: total size of all virtual disks and is a sum of the Allocated metric in Figure 6. In contrast to the datastore Total Capacity metric, there is no indication of how much available capacity there is to add more virtual disks. This is reporting the maximum disk space is available to all VMs assuming no further virtual disks are added.
- Used Capacity: amount of space used, including all snapshots. Used Capacity can be higher than Total Capacity when thin provisioning technologies are used.
- Snapshot/Delta: size of incremental and full snapshot files related to virtual machines that contribute to Used Capacity and Provisioned Capacity.
The following points are wroth considering:
- When the snapshot of a disk is taken, the current state of the disk is retained and a 0 KB delta file is created for future writes. If another snapshot is taken, current delta file will be retained, and the new delta file will be created. The snapshot tree will contain of base file and a sequence of delta files.
- A snapshot file, containing the current state of memory, is also created and can reach the size of memory allocated to the VM.
When is the Virtual Disk View relevant?
If there is a need to understand the utilization of storage currently allocated to virtual machines, the VMware Virtual Machines Storage view provides this information.
A specific case is when migrating to a cloud environment or any infrastructure where the underlying disk storage properties are unknown. In this case, leveraging the Used Capacity is critical as an upper bound for space required to successfully do the migration.