vSphere 5 – Set Round Robin Defaults for EMC Storage

Changes to vSphere 5 syntax has changed the old commands and added new functionality.

To set Symmetrix arrays to a default of Round Robin, for example, you would now use “esxcli storage nmp satp set –default-psp VMW_PSP_RR –satp VMW_SATP_SYMM”

VMware KB1024051 – MSCS vCenter Clustering Unsupported

A newly published VMware KB1024051 article now states that “VMware does not support third party clustering products.”

This means that Microsoft Clustering Service or Veritas Cluster Services are not supported for vCenter 4.x.

Please take this into account when designing mission-critical vCenter deployments.  These customers should be looking at VMware vCenter Server Heartbeat with VMware HA instead of third-party clustering products.

VMware vCenter Server Heartbeat

VMware KB1007566 – Monitor Snapshot Deletion

VMware updated a KB today on how to monitor snapshot deletion.  This is especially useful for large snapshots that “time out” for vCenter because they take more than 5-10 minutes to complete.  I have been using this exact procedure for years and never thought to blog it…for some insane reason.

Read the KB here.

Basically, use the watch command on the *.vmdk files in the VM datastore.  You can use -n to set a refresh interval and -d to show the change in size between refreshes.  However, the most important information is what base .vmdk file is being written to and when.  This touch timestamp will give you more information on which disk is in progress, which has not yet started, and what is already done.

Interpreting ESXTOP files with ESXPlot

Scott Drummonds of vPivot.com posted on ESXPlot, a great tool that is far more efficient and less aggravating than Perfmon for working with ESXTOP output files.

Geoff White has created a badly needed tool here, Perfmon worked but was very time consuming.  Especially with those days or weeks of ESXTOP statistic files.

Extract and use the README.  For Windows, extract all files, navigate to the esxplot.exe in the Bin directory and you are off and running!

Geoff White now has a full project site with all information and releases.  Click here

Labs.vmware.com has now gone live and hosting ESXPlot as well.  Click here

VMware Clustering vCenter Server 4 on MSCS

*NOTICE* As of July 16, 2010 – VMware KB1024051 – has been published stating that “VMware does not support third party clustering products.”  The instructions below should not be used for any production or supported environment.

VIOPS has a good document on how to cluster the vCenter services on MSCS.  Things have changed slightly from 2.5 and though most of the steps will be old news for some, there are a couple changes so be sure to review before proceeding with your Install/Upgrade projects.  It does have a nice section on upgrading existing clusters as well.

Get the document here.

One special note I would point out is running the “jointool.bat” for replicating roles at the end of your cluster config.

VMware KB1003797 – Unable to migrate: Virtual machine is configured to use a device that prevents migration

There seems to be a lot of confusion about vMotion and bus-shared VMDKs.  For the record, and from VMware, it is not recommended nor supported to vMotion your VM between ESX Hosts when using either Virtual or Physical mode bus-sharing.

Please see the KB here.

I know there are plenty that will disagree and have found ways of doing it, but within our environments for our MSCS clusters using Physical mode, we disable DRS and only Cold vMotion once the resources of the cluster have been moved off the Node and the VM is Powered Off.

VMware vSphere Hardening Guide – Draft Rev. B

VMware has released a draft of the vSphere Hardening Guide, and are encouraging feedback from the community.  It is broken into multiple pieces.

Part 1: Introduction

Part 2: Virtual Machines

Part 3: Host

Part 4: vNetwork


Part 5: vCenter

Part 6: COS

*Update*  An official Hardening Guide for vSphere 4 has been released http://communities.vmware.com/docs/DOC-12306










VMware KB1011170 – Checking Zero vs EagerZero Disk Status

VMware has released a KB with a handy vmfstools command to check your VM’s disk status.  If the VMFS output shows a “Z” it is Zerothick with blocks that have not been written to yet.

Read the article here.

This allows you to check the disk status for options such as Fault Tolerance or Microsoft Cluster Services.  The disks must be in an Eagerzerothick format for these options to work.