ESX/ESXi/vCenter Ports

Here is a link to a great PDF diagram outlining all of the network port requirements for your virtual environment. Great resource to have, and your firewall admin will love you! vSphere Network Ports

Read More

Network Ports

ESX vs Hyper-V…VMware vs Microsoft

Like most others out there, I have recently started hearing the ESX/VMware vs Hyper-V/Microsoft questions popping up throughout my day. Check here to get my thoughts!

Read More

Microsoft vs VMware

ESX/ESXi root login over SSH

This is an old one, but worthy of documentation… By default root login over SSH is disabled in ESX 3.X and vSphere 4.X. While it is best practice for security to leave this disabled, from time to time a situation [...]

Read More

putty

ESX – 64GB Memory Limitation

0

Trying to install ESX  3.X on a new server with greater than 64GB of RAM? Then you’ve probably run into a problem. When looking at the host summary in VC, a host with no VMs on it will appear to be utilizing almost 50% of it’s available RAM!  This bug was supposedly fixed via patch, but if you’re running a server that doesn’t have this patch for whatever reason, this might help.

To enable 40 Bit Memory Type Range Registers (MTTR), which is disabled by default, go to the Configuration Tab on your ESX 3.X host. From there navigate to Advanced Settings > VMkernel and find the configuration entry for VMkernel.Boot.force36BitMTRRMask.

Setting

Simply uncheck this setting, and reboot your host.  When the host comes back online, your RAM usage will be accurate. Again, I believe this bug was fixed in 3.5 U2, but I felt it was important to mention the issue in case folks out there aren’t at U4.

I have tested this fix with several Dell R900 hosts without any issue.  If you run into an issue, do the following:

  1. Boot the host into single user mode
  2. Edit  /etc/vmware/esx.conf
  3. Add the following line at the end of the file /vmkernel/force36BitMTRRMask = “TRUE”
  4. Run esxcfg-boot -b
  5. Reboot the host

ESX – Options for setting a static MAC in a VM

0

Along the way you might run into a P2V (or need to build a new VM) that will host an application which has a license attached to it’s NIC. In order to do this, you generally need to provide the software vendor the MAC address to your VM. That being said, it’s important to ensure that the MAC address never changes. This can be accomplished 2 ways.

The first option is by setting a new static MAC  at the VM level. VMware has made this possible by blocking out a range that can be used for static addresses: 00:50:56:00:00:00 to 00:50:56:3F:FF:FF  

The first three octets of any MAC address are a unique code assigned to each NIC vendor.  This is also known as an Organizationally Unique Identifier or OUI.  VMware uses an OUI of 00:50:56 . The last 3 octets can be any combination as long as that combination is unique on your network. If using a VMware static MAC, you must also stay inside the range mentioned above.

  1. Power off the VM and then remove it from the Virtual Center inventory. If you do not remove the VM from the Virtual Center inventory, VC will overwrite the changes you’ll make below when you power the VM back on. This step is necessary!
  2. Edit the VM .vmx file (either via the datastore browser, or via an SSH session on your host) and locate the already pre-existing generated MAC address:
    EthernetX.addressType =
    and
    EthernetX.generatedAddress =
    where X is the number of your ethernet adapter
  3. Change the value for EthernetX.addressType from ”vpx” to “static”
  4. Change EthernetX.GeneratedAddress to EthernetX.Address
  5. Change the current MAC to a new static MAC address using the allowed range I mentioned above
  6. In Virtual Center, select your ESX host and browse the datastore the VM your working on is located in
  7. Right-click on the .vmx file, and select “Add to inventory”
  8. Power on the VM. Issue the command “ipconfig /all” and verify the manually assigned MAC address

Tip: Run-Virtual.com has published a pretty nice tool that allows you to set the MAC address of your VMs. I’ve never personally used this tool, but have read good things about it.

The second option is to set the MAC address from within the guest operating system. This would be necessary if you were doing a P2V in which the MAC of the physical NIC is attached to your application license, and you simply can’t change to a new static MAC. In the case of  Windows Server 2003, changing the MAC address is as simple as modifying the “networkaddress” field within the properties for the NIC. Changing this address will override any setting at the VM level. Keep in mind that if you are going to use the same MAC as that of a physical adapter, that adapter can no longer be used in your environment at the same time.

VMware SRM – Datastore Naming

0

When failing over from a production site to a DR site, the datastores that are mounted on the ESX hosts at your DR site are named using the “snap-xxxxxxx-NAME” naming convention by default.

srm-snap-mount1

Per this thread, this was not this way in the SRM beta, but was changed for the GA release so folks would be able to easily identify snapshots.  You can disable the renaming process by doing the following…

1. Locate vmware-dr.xml in the C:\Program Files\Site Recovery Manager\Config directory
2. Change <fixRecoveredDatastoreNames>false</fixRecoveredDatastoreNames> to be <fixRecoveredDatastoreNames>true</fixRecoveredDatastoreNames>

Thanks to the folks on VMware Communties for the great tip.

VMware SRM – My initial thoughts

0

Recently, I had an opportunity to install and configure SRM in a test lab, as part of a project related to a datacenter move. VMware Site Recovery Manager is THE tool to utilize when you’re looking to create an automated solution for disaster recovery. One great thing about SRM is that it can be used for a move project, or for a DR project, so those who are planning  a move of their virtual environment actually will get double the value out of their licensing purchases.

Here are the details of the environment I’ve setup…

  1. 2 ESX hosts running ESX 3.5U4
  2. 2 VC servers (each with one of the above hosts connected to it)
  3. 2 EMC Clariion Storage arrays (one zoned to each host)
  4. Powerpath installed on the CX arrays, SRM licensing on the VC servers

This effectively creates me a production site, and a DR/move site.  Using a great install/configure document at VIOPS: Proven Practice site , the setup and configuration of Mirrorview (replication for the Clariion storage array), the Clariion Storage Replication adapter, and SRM were a breeze. I definitely learned a few “gotchas” along the way, and plan to share them in future posts. One thing I’ll mention now is that SRM is only as good as the replication happening in the background. VMware works very closely with the various different storage vendors to develop the replication adapters that actually get SRM and the storage arrays communicating. Consult with your storage vendor beforehand to ensure that your array is supported with SRM. Being that I was working with EMC gear, I did not run into any support issues.

Once everything was setup and working I was able to create a quick recovery plan, test the plan, and then run the plan, all in a matter of minutes. After working through a few issues, SRM performed as expected and migrated my VMs from a production site to a DR site. 

I would encourage anyone who hasn’t yet tried SRM out to go to www.vmware.com and download the trial licensing. Set it up in your test environment and gauge whether or not it has any value for you and your organization as a DR enabler. As always, the VMware Communities is a great resource for anyone setting up SRM who might run into an issue. Lastly, keep in mind that SRM is not yet supported when running vSphere. A point release is due out later this year that will add support.

vSphere Fault Tolerance Simulation

0

vmware.info has posted a really nice flash simulator for setting up Fault Tolerance in vSphere. This simulation gives you a great idea of how to configure and test the Fault Tolerance feature if you aren’t running vSphere in your lab. I’ll go more into what Fault Tolerance is, and how to use it in future posts.  Until then,  Check out the simulation!

vmwareft

Host requirements:

  1. CPUs: Only recent HV-compatible processors (AMD Barcelona+, Intel Harpertown+), processors must be the same family
  2. All hosts must be running the same build of VMware ESX
  3. Storage: shared storage (FC, iSCSI, or NAS)
  4. Hosts must be in an HA-enabled cluster
  5. Separate VMotion NIC and FT logging NIC, each Gigabit Ethernet (10GB recommended)
  6. System BIOS must be at mnimum version
  7. Hardware virtualization (Intel VT for Intel Procs) must be enabled

VM requirements:

  1. VMs must be single-processor (no vSMP)
  2. All VM disks must be “thick” (fully-allocated)
  3. No non-replayable devices (USB, sound, physical CD-ROM, physical floppy, physical RDMs)
  4. Make sure paravirtualization is not enabled by default

Per processor licensing for virtual machines

0

I read a great post on Friday over at Yellow-Bricks  (which by the way, is one of my favorite virtulization blogs to follow). I think this “feature” is great, and can’t wait to hear if it is fully supported or not. Surely, this could save folks a ton in licensing if they’re using products that are licensed per CPU socket. 

“Some vendors license their application per processor, also in a virtualized environment. So if your VM has 4 vCPU’s your vendor will want you to buy a 4 processor license for the application. But you can avoid this by telling the VM that it has cores instead of processors. In others words, instead of having 4 processors you would have 1 processor with 4 cores:

  1. Power off the VM
  2. Right click on the VM and select “Edit Settings…”
  3. Select the “Options” tab
  4. Click on “General” (in the “Advanced” options section)
  5. Click “Configuration Parameters…” (in the pane on the right)
  6. Click “Add Row”
  7. Enter “cpuid.coresPerSocket” in the “Name” column
  8. Enter a value (try 2, 4, or 8 in the “Value” column
  9. Click “OK”
  10. Power on the VM”

Click here to read the full post!

Page 6 of 7« First...34567