Information Technology Reference
In-Depth Information
information will provide a foundation upon which we will build later in the chapter. We'll start
with a look at the vSphere Virtual Machine File System, a key technology found in practically
every vSphere deployment.
Examining the vSphere Virtual Machine File System
The vSphere Virtual Machine File System (VMFS) is a common coni guration option for many
vSphere deployments. It's similar to NTFS for Windows Server and ext3 for Linux. Like these
i le systems, it is native; it's included with vSphere and operates on top of block storage objects.
If you're leveraging any form of block storage, you're using VMFS.
The purpose of VMFS is to simplify the storage environment. It would clearly be difi cult to
scale a virtual environment if each VM directly accessed its own storage rather than storing the
set of i les on a shared volume. VMFS creates a shared storage pool that is used for one or
more VMs.
While similar to NTFS and ext3, VMFS differs from these common i le systems in several
important ways:
◆
It was designed to be a clustered i le system from its inception; neither NTFS nor ext3 is a
clustered i le system. Unlike many clustered i le systems, it is simple and easy to use.
VMFS's simplicity is derived from its simple and transparent distributed locking mecha-
nism. This is generally much simpler than traditional clustered i le systems with network
cluster lock managers.
◆
VMFS enables simple direct-to-disk, steady-state I/O that results in high throughput at a
low CPU overhead for the ESXi hosts.
◆
Locking is handled using metadata in a hidden section of the i le system, as illustrated
in Figure 6.17. The metadata portion of the i le system contains critical information in the
form of on-disk lock structures (i les), such as which ESXi host is the current owner of a
given VM, ensuring that there is no contention or corruption of the VM.
◆
Depending on the storage array's support for VAAI, when these on-disk lock structures are
updated, the ESXi host performing the update momentarily locks the LUN using a nonper-
sistent SCSI lock (SCSI Reserve/Reset commands). This operation is completely transpar-
ent to the vSphere administrator.
◆
These metadata updates do
not
occur during normal read/write I/O operations and do not
represent a fundamental scaling limit when compared with more traditional i le systems.
◆
During the metadata updates, there is minimal impact to the production I/O (covered in a
VMware white paper at
www.vmware.com/resources/techresources/1059). Th
is impact
is negligible to the ESXi host holding the SCSI lock but more pronounced on the other hosts
accessing the same VMFS datastore.
◆
These metadata updates include, but are not limited to the following:
◆
The creation of a i le in the VMFS datastore (powering on a VM, creating/deleting a
VM, or taking a snapshot, for example)
◆
◆
Actions that change the ESXi host that owns a VM (vMotion and vSphere HA)
