Configuring Storage Area Networks

VMware ESX Server 2.1
Features | Documentation | Knowledge Base | Discussion Forums

Use the Storage Management option to manage your storage area network and attached storage devices for your ESX Server system and its virtual machines.

Because the disks on the SANs can potentially be accessed by multiple ESX Server computers, there are some configuration issues that are unique to SANs.

For more information about SANs, see Using Storage Area Networks with ESX Server.

Note: Be sure that only one ESX Server system has access to the SAN while you are using the VMware Management Interface to configure it by formatting the VMFS-2 volumes. After you have finished the configuration, be sure that all partitions on the shared disk are set for public or shared access for access by multiple ESX Servers (see VMFS Accessibility).

Configuring Storage: Disk Partitions and File Systems

The Disks and LUNS window allows you to view and modify the partitions and file systems on your disks. Create disk partitions that use the VMFS file system, suitable for storing disks for virtual machines. You can also edit, label and remove existing partitions. When you edit a VMFS partition, you can change the volume label, maximum file size, access mode and whether you want to span the partition.

Creating a Disk Partition

You can use any existing free space on your VMFS volumes to create new disk partitions. For background on how SCSI devices are identified, see Determining SCSI Target IDs.

Note: You cannot change any partitions set up when you installed ESX Server. These include any volumes with a Linux file system or that are used for Linux swap space.

If a core dump file does not exist on the disk, you are offered to create one. Creating a new volume consumes all the free space remaining on a disk. ESX Server determines the optimum setting for the maximum file size based on the volume's file system.

Caution: Do not place your core dump partition on a SAN disk. If the SAN cable is removed, ESX Server becomes unstable.

To create a new partition, complete the following steps.

In the Disks and LUNs window, click Create Volume. You are asked how you prefer to configure the disk.
Click Typical.
If it does not exist, you are asked if you want to create a core dump partition.

The core dump partition stores information generated if the VMkernel crashes. The core dump information is important in debugging any problems with the VMkernel.
The rest of the disk or array is used as a VMFS partition, where you store virtual machine disk files. The VMFS partition provides high-performance access to the virtual machine's files — essentially the same performance you would get if the virtual machine were installed on a raw SCSI partition.
Note: Keep in mind that only four primary partitions can exist on a drive. If you have an extended partition (to contain logical partitions), that counts as one of your four primary partitions.
Click Yes to create the core dump partition. ESX Server also creates the VMFS partition.

After you create the partition, you can add a volume label, determine access mode and the maximum file size and span the disk with any public extents. For more information about access modes, see VMFS Accessibility.

Editing a Disk Partition

Select a partition to edit and click Edit.

ESX Server 2.1 includes a new file system, VMFS version 2 (VMFS-2). If this partition is formatted for VMFS-1, you can convert it to VMFS-2. See File System Management on SCSI Disks and RAID for detailed information on this file system.

The changes you can make to the partition may include (certain partitions do not allow you to change all of the following):

Setting the volume's type.
Changing the name of the volume label.
Setting the volume's access mode.
Setting the volume's maximum file size.

Setting the Volume's Access Mode

There are two modes for accessing VMFS volumes: public or shared.

Public mode is the default mode for ESX Server. VMware recommends you use this mode.
With a public VMFS version 1 (VMFS-1) volume, multiple ESX Server computers have the ability to access the VMware ESX Server file system, as long as the VMFS volume is on a shared storage system (for example, a VMFS on a storage area network). However, only one ESX Server can access the VMFS volume at a time.
With a public VMFS version 2 (VMFS-2) volume, multiple ESX Server computers can access the VMware ESX Server file system concurrently. VMware ESX Server file systems with a public mode have automatic locking to ensure file system consistency.
Shared mode is used for a VMFS volume that is used for failover-based clustering among virtual machines on the same or different ESX Servers.
Note: If you plan to change the accessibility mode for a VMFS volume, you must deactivate the swap file if it exists. To deactivate the swap file, see Configuring a Swap File.

Changing the Maximum Size of a File Allowed by VMFS

If you plan to create virtual machines with virtual disks larger than the default maximum size of 144GB, change the value in the Max File Size field.

Spanning a VMFS volume.

You can only span VMFS-2 volumes. Spanning a volume allows the volume to comprise multiple VMFS disk partitions. Each disk or partition to which this volume is spanned is called an extent. In effect, this creates a single volume that is larger than would be possible from one partition. Also, in the spanned volume or extent, you cannot change the maximum size of files.

Once you span a volume, you cannot remove the volume if it is spanned or if it spans other volumes. To span to another volume, check the box next to that volume label.

Caution: Any data on the extent is lost when the VMFS volume spans to it, so it is a good idea to span to newly created partitions.

Converting a Partition to VMFS-2

To convert the partition to VMFS-2, click the Convert to VMFS-2 link. In order to convert the file system, you must deactivate the swap file if it exists. To deactivate the swap file, see Configuring a Swap File.

Caution: Metadata on VMFS-2 volumes utilize more space than metadata on VMFS-1 volumes. To successfully convert a file partition, you may need to move files to allow for more disk space.

Removing a Disk Partition

To remove the partition, click Remove. You are asked to confirm that you want to remove the partition. To delete certain partitions, you must click Edit, then Remove.

If the volume is spanned to other volumes, you cannot remove it.

For more information, see File System Management on SCSI Disks and RAID.

Convert the partition to VMFS-2. Click the Convert to VMFS-2 link. In order to convert the file system, you must deactivate the swap partition if it exists. To deactivate the partition, see Configuring a Swap File.
Removing the partition.
Click Remove. You are asked to confirm that you want to remove the partition. You cannot remove a volume that is spanned.

Adapter Bindings

This Adapter Bindings view displays the World Wide Port Names bound to each Fibre Channel host bus adapter (HBA) in the system. You can also view the persistent binding status for each HBA. With persistent bindings, ESX Server assigns specific target IDs to specific SCSI devices. This target ID association is retained from reboot to reboot unless changed by you.

Persistent bindings are particularly useful if you are using raw disks with ESX Server. A raw disk is directly mapped to a physical disk drive on your storage area network (SAN). ESX Server directly accesses the data on this disk as a raw device (and not as a file on a VMFS volume).

Viewing Failover Paths Connections

The Failover Paths page allows you to review the current state of paths between your system and SAN LUNs. Multipathing support allows your system to maintain a constant connection between the server machine and the storage device in case of the failure of a host bus adapter (HBA), switch, storage controller, or a Fibre Channel cable.

For each SAN Logical Unit Number (LUN), this page displays the available paths and the preferred path. By default, ESX Server selects the last path used to access a LUN.

The failover paths show the adapter, target, LUN and the SAN target for the LUN. Each SAN target is identified by its World Wide Port Name.

The status of each path is indicated by a symbol that corresponds to its current status.

— This indicates that the path is active and data is being transferred successfully.

— This indicates that the path is set to disabled and is available for activation.

— This indicates that the path should be active, but the software cannot connect to the LUN through this path.

If you have configured a LUN to use a preferred path, that path will be identified with the label Preferred after the SAN Target listing.

Configuring Failover Policies

The failover paths edit feature allows you to configure the policy for transferring LUN access from one path to another.

To edit the failover policy for a LUN:

From the Failover Paths page, click Edit. The configuration page appears and displays information about the current state of the paths and failover policy options.
Choose one of the following failover policies:
Fixed — always use the preferred path when available
Most Recently Used — always use the last active path
If you select Fixed, next choose the preferred path by selecting Preferred in the Adapter icon pulldown menu for that path.
Click OK to save your settings and return to the Failover Paths page. The name of the failover policy appears next to each SAN LUN in the failover paths list.

For more information on failover policies, see Setting Your Multipathing Policy for a LUN.

Configuring Failover Paths

You can also enable or disable individual failover paths by changing their status in the Adapter icon pulldown menu.

Prev Contents Last Next

Support Resources