VM /ETC

Posts Tagged ‘vmfs’

It’s inevitable. If you are a virtual infrastructure administrator, architect, or blogger you will eventually talk about VMware VMFS and virtual disk alignment. I’m not going to try to explain the concepts in this post. My goal is to discuss whether to align VMware virtual disks (VMDKs) and what options are available for performing the alignment.

Bottom line is that if you are deploying new VMs from a template you should do so from an aligned master, but VMs created from physical to virtual migrations (P2Vs) present some practical challenges and don’t have to be treated as urgently.

Recommendations

First things first. My interpretation of VMware’s stance on alignment is that they do not advocate aligning every P2V-ed VM to the VMFS due to the administrative effort, disk space requirements, potential VM down time, and in most cases a nominal VM performance increase. On the other hand, it is crystal clear that VMware does recommend always formatting VMFS LUNs with the VI Client and always taking the time to align Gold Image templates so new servers deployed will automatically be aligned.

For reference here’s VMware’s PDF on the topic:

I’ll admit my perspective as a consultant may be a bit different than the normal VI admin. After all, I am mostly involved in the Read the rest of this entry »

“Should I use VMFS or RDMs (Raw Device Mappings) for my VMware ESX virtual machines (VMs)?” It’s a popular discussion not just for new virtual infrastructure (VI) planning but for existing environments as well. In my experience, Many VMware ESX virtual machines using RDMs were created because of a perceived boost in VM and application performance. Somehow, the wrong impression became widespread: the proprietary VMware VMFS file system created too much I/O overhead and slowed down the VM performance. Although the VMFS reputation has been publicly repaired again, there are still regular challenges explaining this virtual disk myth to VMware administrators.

There are definite pros and cons for using both VMFS and RDMs. This post suggests the 3 most common practical reasons (in my opinion) to use RDMs. That is, reasons that benefit VI administrators, leverage a VM’s native operating system,  or take advantage of technologies and process designed for physical environments. I also provide some links to a few other blogs that have recently explored the same RDM vs VMFS topic as supporting examples.

First, if proof is still needed that performance should not be a factor in the VMFS versus RDM decision, check out VMware’s own PDF titled Performance Characteristics of VMFS and RDM.

So what are some practical reason for using RDMs? Read the rest of this entry »

Fluid Operations, owner of the eCloudManager Project, has open sourced it’s VMFS Driver Project. The driver is actually an application that can run on Linux, Windows or any operating system that supports Java. It apparently provides read only access to VMware VMFS volumes.

The VMFS driver has been developed for use in fluid Operations’ eCloudManager Storage Edition which is described as:

“… the industry’s first solution which allows for managing CPU and storage virtualization through a single pane of glass. This solution allows high-performance applications to be rolled out on an enterprise cloud utilizing multi-vendor server virtualization and storage technologies. This first release orchestrates VMware Virtual Infrastructure as server virtualization and NetApp filer technology as storage.”

eCloudManager is in beta, but is available both as a plugin for VirtualCenter or as a web interface.

The implications of the open source availability of the VMFS driver are huge in my opinion. For example, a VMFS driver could allow the creation of a liveCD that would be able to access virtual machines on a failed ESX host. Those stranded VMs could then be transferred to another working ESX host. There are also numerous other VM backup, disk analysis, and monitoring possibilities from a non VMware server.

Here’s some more information from the fluid Operations Community Page:

“This driver enables read-only access to files and folders on partitions formatted in the Virtual Machine File System (VMFS) by VMware. VMFS is a clustered file system that is used by the VMware ESX hosts to store virtual machines and virtual disk files. The VMFS driver is developed and maintained by fluid Operations and is included in upcoming releases of the eCloudManager product, where it is used to allow enhanced features like offloaded backups of virtual machines hosted on VMware ESX hosts. The VMFS driver comes with a command line interface (CLI) to mount and analyze VMFS volumes. The VMFS driver was tested on Linux and Windows based hosts, but should work on any platform that supports Java.”

The VMFS Driver Java application is available for download via it’s Google Code Project page. There is also a FAQ with the following additional information (and more – go check it out) about the VMFS driver. Read the rest of this entry »

If you are considering virtualizing SQL, check out a great post from vm0.blogspot.com. Running SQL Server on ESX makes some great suggestions for planning  CPU, RAM, and Disk I/O to allow for the highest performance possible of your database virtual machines (VM).

The section on disk I/O made me think – specifically about whether creating RDMs to raw disks was really a better choice than carefully planning VMFS LUNs. There is just too much convenience with using VMFS in my opinion, and I’ve never really been convinced that any report or testing has proven RDMs offer better performance than carefully planned VMFS. So, instead of a long comment on the vm0 post I decided to blog about it here at VM /ETC.

Be sure to read the entire post for the RAM and CPU recommendations, but here is the section that inspired me: Read the rest of this entry »

John Troyer has announced that the VMware website has been updated to reflect new product line names first introduced back in September at VMworld 2008. In the VMTN post Do they smell as sweet? New product line names: vCenter, View, John explains that the VirtualCenter and VDM products have are now referred to as vCenter and VMware View. Several VC add ons such as Update Manager and Converter are included along with Site Recovery Manager (SRM), Lab Manager, Lifecycle Manager, and even the file system VMFS now have slightly new names.

John provides a complete list of all products impacted by the name change which I have copied here. Read the rest of this entry »

New in the latest version of ESX 3.5 is an experimental Service Console script called the VMDK Recovery Tool. Since it’s Console based, it’s not available in ESXi. The Update 3 Release Notes say the following about it:

Experimental Support for the VMDK Recovery Tool — This release adds support for the VMDK Recovery tool, a script intended to help customers to recover VMFS/vmdk data stores from accidental deletion of VMFS/vmdk data store or physical disk corruption. For more information, see VMDK Recovery Tool (ESX 3.5 Update 3) ( KB 1007243).

Use the VMDK Recovery Tool to:

Create a “block list” of your VMDK files. The block list file is a plain text file that keeps a record of all VMDKs of the virtual machines.

Recover VMDK files if they become deleted, or if the VMFS datastore gets deleted or corrupted.

Undelete VMDKs? Recover corrupted or deleted VMFS data stores? Reading KB 1007243 a little further reveals you can back up and restore VMs too – either specify which one you want to back up or pick all the registered VMs on the host. The script asks you to specify the directory path to use for saving backups and the destination path for restoring from backups.

Sounds to me like this script could become a “super VCB” without the extra server requirements? It seems much simpler anyways. It will be interesting to see how this script is developed and exactly how much disk space the block list and backup files consume.

Based on best practices, this post is a “formula” for planning and sizing SAN storage for maximum VMFS performance. This is intended to be for all protocols where VMFS volumes are utilized ( FC, iSCSI ) and conservative enough to allow general sizing estimates while ensuring high performance of the running VMs.

WARNING: The storage design resulting from following these recommendations will not be the most cost effective solution. As storage performance generally requires the most spindles (disks) possible and this formula requires as many dedicated LUNs as possible, the cost for storage will be maximized. More often than not, compromises between performance and cost have to be reached that keep the design within the expected budget.

This post is split into 2 sections. The first section lists the VMFS Maximum Performance Rules while the second section uses a 25 server example to walk through the design rules. Read the rest of this entry »