NetApp Tech OnTap
     

Eliminate VMware Storage Provisioning Headaches

Storage management places a burden on more than just traditional storage administrators. Today many VMware (or VI) administrators find that they are spending more and more time on storage operations and storage-related issues.

For example, the task of bringing storage online is a multistep process that requires the coordination of information and tasks between storage and VMware admin teams. Unfortunately for VI admins, they are responsible for a large portion of the tasks related to this fundamental operation.

VMware storage provisioning processes. Note that most tasks are done by VMware admins.

Figure 1) VMware storage provisioning processes. Note that most tasks are done by VMware admins.

NetApp engineering has worked with VMware engineering to simplify storage management. The result is our plug-ins for VMware® vCenter™ Server: the Rapid Cloning Utility (RCU) and the Virtual Storage Console (VSC). When used together they streamline storage operations for both storage admins and VI admins.

If you’ve been reading Tech OnTap in recent months, you may be familiar with these utilities. (Refer to A Quantum Leap in Virtual Management for more.) In this article, I talk about the overall operational impact of these plug-ins, rather than delve into the specifics of each tool. In doing so I introduce a new model for VMware storage management—one that significantly simplifies storage provisioning, cloning, monitoring, and management.

The New Model for Storage Admins

In the new model, storage admins continue to be responsible for configuring physical storage systems, providing data protection, and managing overall utilization, as always. Once the physical architecture is deployed, however, all that a NetApp storage admin supporting VMware has to do is provision pools of storage resources (aggregates, FlexVol® volumes, and storage network interfaces) that VI admins can then configure as needed for use with VMware virtual machines.

This model significantly simplifies storage admin tasks while allowing the VMware admin to directly provision and manage datastores and the storage constructs associated with them (LUN masking, storage I/O path management, etc.) using the resources assigned by the storage admin.

To assign resources, the storage admin logs in to vCenter and opens the RCU configuration panel. This is where he or she may assign a controller or controllers and storage resources. Once a resource has been assigned it is used exclusively by virtual infrastructure. Resources not assigned are ignored and are inaccessible to the virtual infrastructure.

Assigning storage resources to VMware using RCU.

Figure 2) Assigning storage resources to VMware using RCU.

To prevent further changes the storage administrator has the option to lock or restrict the ability to assign additional resources. (Taking this step is a recommended best practice.) Checking a box followed by entering a user name and password is all that’s needed to perform the securing process.

That's it! That’s all a storage admin needs to do to provide storage for a VMware environment. There are no LUNs, no LUN masking, no NFS exports, no multipathing.

The New Model for VMware Admins

Now, let’s look at the impact the new model has on VMware admins in terms of provisioning and managing datastores.

Provisioning Datastores from Assigned Resource Pools
Within vCenter a VMware admin can right-click on an ESX/ESXi host, cluster, or data center and provision a datastore for that unit via the NetApp option in the menu. The datastore can be FC, FCoE, iSCSI, or NFS and will be configured from one of the resource pools established by the storage admin. The plug-in will handle path selection, provide load balancing through VMware native multipathing policies, secure the storage target, and enable thin provisioning and data deduplication. While the VMware admin benefits from automation and an on-demand provisioning process, the overall environment also benefits from the consistent implementation of NetApp best practices.

Provisioning datastores from vCenter

Figure 3) Provisioning datastores from vCenter.

Managing Datastores
Ever provision storage and later wish you could modify it, perhaps when a datastore is nearing its capacity limit? With NetApp you have additional choices in the areas of dynamic nondisruptive resizing of datastores. (Note: Only NFS datastores can be shrunk.) To change the capacity or storage efficiency settings, you simply select the datastore, right-click, and choose the appropriate option. It’s that easy.

Instant Provisioning of VMs
Development of the NetApp plug-ins started with a number of us at NetApp and VMware looking at ways to reduce storage costs for virtual desktop deployments using NetApp FlexClone® cloning technology for cloning individual files. Until that time, FlexClone was limited to use on entire LUNs and FlexVol volumes.

Today, this technology has evolved to the point where you can deploy a single VM, multiple VMs, or a pool of virtual desktops almost instantly, without consuming any additional storage in the process. A VMware admin can deploy space-efficient clones by simply selecting a running or shut-down VM, template, or vApp as a starting point.

Unlike some cloning technologies, NetApp clones are permanent, high-performance VMs that can be treated in the same manner as any other VM. There are no restrictions. Perhaps as a result, NetApp FlexClone technology has been integrated into vCenter, View Manager, XenDesktop, and Quest vWorkspace.

Audit and Automate VMware Storage Settings
By selecting the NetApp tab in vCenter, a VMware admin can view the NetApp storage systems configured for use in the virtual infrastructure. Here you can identify the ESX/ESXi hosts connected to each storage system and audit the settings related to FC, FCoE, iSCSI, and NFS to verify that they are set to values defined in NetApp best- practices guides TR-3428 (VI3) and TR-3749 (vSphere). Should these settings require updating, you can select one or multiple hosts and execute a nondisruptive update to the storage settings.

This audit process can be run at any time without disruption to production, providing the ability to enable optimal uptime as your environment grows. These storage settings are currently not covered by VMware host profiles, so the NetApp plug-ins extend the completeness of an automated deployment process. The ability to update host settings is limited to vSphereâ„¢ hosts. VI3 hosts lack the API calls to automate this process. Should a system be identified as being out of compliance, the storage settings can be changed manually as outlined in TR-3428.

Report on Storage Details and Utilization
Also available from the NetApp tab is the ability to report on storage utilization for SAN- and NAS-based datastores. When you select a datastore, a large amount of details on the underlying storage object—such as LUN serial number, igroup, ALUA enablement, deduplication savings, etc.—can be reviewed.

A key benefit is the ability to report on storage utilization through the various storage layers beginning at the datastore and ending at the NetApp aggregate (a collection of RAID-protected disks). The value of NetApp integration with vCenter becomes obvious when storage-saving technologies such as data deduplication, FlexClone, or thin provisioning have been enabled.

Report Storage Faults
The NetApp plug-in also gives VMware admins feedback on the health of the storage controllers themselves. The ability to report the health of the physical infrastructure can reduce the time it takes to address a problem.

Enable Optimal I/O Settings Within VMs
Another component of the NetApp plug-in provides the ability to audit and adjust settings within a VM to enable optimal I/O. The first set of tools includes scripts that can be run from within VMs or applied to VM templates to set local SCSI settings within the guest operating system.

Second, MBRscan and MBRalign combine to audit and correct the partitions and file systems within a VM to make sure that VM I/O operations are aligned to storage system block boundaries. The problem of virtual machine disk alignment is not unique to VMware, nor is it unique to NetApp storage. This problem can occur in any virtual environment on any storage platform. (I wrote a recent post on this issue for my Virtual Storage Guy blog with help from Duncan Epping of VMware.)

This problem occurs because, by default, many guest operating systems, including Windows® 2000 and 2003 and various Linux® distributions, start the first primary partition at sector (logical block) 63. This behavior leads to misaligned file systems because the partition does not begin at a block boundary. As a result, every time the virtual machine wants to read a range of data blocks, an additional block has to be read. This results in unnecessary load on the storage controller. You can learn more about the problem and how to correct it in TR-3747. This document has been reviewed and approved by VMware, Microsoft, Citrix, and NetApp.

Table 1) Storage system and VMware plug-in requirements.

• NetApp FAS, V-Series, or IBM N series running Data ONTAP® 7.3.1.1 or later
• RCU 3.0 vCenter plug-in installed
• VSC 1.0 vCenter plug-in installed
• VMware ESX/ESXi 3.5 or 4.0 (and subsequent updates)
• VMware vCenter server 2.5 for RCU or 4.0 for RCU and VSC
• Copy of best-practices guide (TR-3749 for vSphere, TR-4328 for VI3)

Conclusion

The advantages created by the use of NetApp storage and our vCenter plug-ins are substantial. The original multistep provisioning model presented in the introduction (see Figure 1) is effectively reduced to a simple two-step model.

New VMware provisioning model made possible by RCU and VSC.

Figure 4) New VMware provisioning model made possible by RCU and VSC.

Working from within vCenter, both storage admins and VMware admins can complete tasks faster with fewer complications and much more control.

NetApp Coommunity
 Got opinions about VMware storage provisioning?

Ask questions, exchange ideas, and share your thoughts
online in NetApp Communities.


Vaughn Stewart

Vaughn Stewart
Virtualization Evangelist
NetApp

Vaughn is the virtualization evangelist for NetApp and a member of the Virtual Solutions Engineering team. He was awarded a vExpert for 2009 by VMware. Vaughn is the coauthor of several white papers on integrating VMware technologies on NetApp systems, and regularly publishes the blog The Virtual Storage Guy. Look for his new book on architecting storage for cloud deployments later this year.

 
Explore