Aug 012011

The hottest topic after the launch of vSphere 5.0 has unfortunately not been about the awesome technology features announced like Auto Deploy, Storage DRS, Fault Domain Manager etc etc.

I was around in Cannes VMworld 09 when vSphere 4.0 was released and all the talk seemed to be about licensing then also, so this controversy does not come as a huge surprise to me.

I am a little disappointed however that some people I know to be extremely clever are failing to see the bigger picture. I’m not going to go into all that now as I think it has been blogged to death. Some good articles are by Rynard Spies who speak about the downside of new licencing compared to vSphere 4.0 and Aaron Delps, speaking about the bigger picture of TCO and how the VMwares new licencing model fits.

Personally the new licencing model does not seem all that bad to me. Those who know me will know that I love a good analogy, so here I go….

I believe comparing the old vSphere 4.x licencing model to 5.0 is seriously floored. Lets use a car as an example:

A Ford Focus Hatchback (05-11 model) cost between £11,695 – £20,845 brand new. a Ford Focus Hatchback (‘11 onwards) costs (£16,000 – £24,000). Now we all know why we are paying more, it’s a better looking car, it has better safety features, it has more toys, it is more efficient on MPG but most critically the development of the car. Even though it is at a high-level it is more or less the same car, no-one complains that the price has gone up considerably.

When VMware went from VI3.x to vSphere 4.0 they didn’t actually put the price up to reflect all the new toys and development costs.

So why have they increased the costs now? Lets look at the industry, even though there are the ‘Scale Out’ gang who blindly believe that it is the best model, the hardware vendors are producing the most powerful servers that we have seen and are managing to pack them into more and more reduced CAB space. Added that CPUs will soon be going in to the high teens and 20’s means that VMware old licencing model would be totally financially unfeasible. If everyone went for 4.x Ent Plus then with all the factors I mentioned above VMware would be selling less and less licences and make less money!!

So, do I think there could be improvements with the licencing model VMware have produced for vSphere 5.0? Yes, I would like to see the vRam limit up to 96GB for Ent Plus (rumours say this might happen). But do I think switching to vRAM was the correct decision? Yes, it makes total sense to me when looking at VMwares long term goal to remain the market leader.

Lets stop moaning about the licencing model and start talking about the product its self.

Post to Twitter

Jul 122011

Bye Bye ESX and the Service Console!! The long time talked about removal of the service console from the vSphere product set means vSphere 5.0 is now a converged platform offering (ESXi).

So what platform enhancements are in the vSphere 5.0 release?

Image Profiles – An image profile allows for the create of bespoke ESXi images. Common issues in previous version of ESX meant that existing ESXi images could have missing drivers and no plugins like nk1v. It describes all the components in an ESXi instillation image, but not configuration. The Image Builder is a set of command line utilities (powercli version). Image Profiles are kept in a depot, a repository (library of images and VIBs)

What is Auto-Deploy?

ESXi is downloaded and ran from memory. VMware vSphere Auto Deploy virtual appliance loads the hypervisor software onto host as it is booting. It uses the image profiles obtaining from the depot and boot it via PXE. Combining the features of host profiles, Image Builder, and PXE, VMware vSphere Auto Deploy simplifies the task of managing ESXi installation and upgrade for hundreds of machines. New hosts are automatically provisioned based on rules defined by the user. Rebuilding a server to a clean slate is as simple as a reboot and the physical memory is cleared. To move between ESXi versions, you update a rule using the Auto Deploy PowerCLI and perform a test compliance and repair operation.

Auto-Deploy allows for rapid provisioning and re-provisioning of new hosts and enables simplified updates.image

The big bonus for Auto-Deploy is no local storage for host hypervisors. The same goes for Boot from SAN.


Recipe for Auto-Deploy:

Custom ESXi image + network boot machine + auto deploy virtual appliance + configuration + host profiles + runtime state information + vCenter server + recording of events + syslog netdump servers


Used to be IPtables – not anymore. Service aware but not state of host. 3rd party components.

Can add rules thorough GUI or xml format.

New Virtual Machine Capabilities

VM scaling:

Hardware version 8. 32 virtual cpu per VM, 1tb ram per VM.

    Support for up to 512 virtual machines. vSphere 5.0 supports up to 512 virtual machines totalling a maximum of 2048 virtual CPUs per host.

    Support for larger systems. vSphere 5.0 supports systems with up to 160 logical CPUs and up to 2TB RAM.

    other new features – UI for multi-core virtual CPUs, extended VMware tools compatibility, support for mac os server.

    • New Virtual machine capabilities. ESXi 5.0 introduces a new generation of virtual hardware with virtual machine hardware version 8, which includes the following new features:
      • 32-way virtual SMP. ESXi 5.0 supports virtual machines with up to 32 virtual CPUs, which lets you run larger CPU-intensive workloads on the VMware ESXi platform.
      • 1TB virtual machine RAM. You can assign up to 1TB of RAM to ESXi 5.0 virtual machines.
      • Nonhardware accelerated 3D graphics for Windows Aero support. ESXi 5.0 supports 3D graphics to run Windows Aero and Basic 3D applications in virtual machines.
      • USB 3.0 device support. ESXi 5.0 features support for USB 3.0 devices in virtual machines with Linux guest operating systems. USB 3.0 devices attached to the client computer running the vSphere Web Client or the vSphere Client can be connected to a virtual machine and accessed within it. USB 3.0 devices connected to the ESXi host are not supported at this time.
      • UEFI virtual BIOS. Virtual machines running on ESXi 5.0 can boot from and use the Unified Extended Firmware Interface (UEFI).

    Other new features – UI for multi-core virtual CPUs, extended VMware tools compatibility, support for mac OS server.


    Storage DRS (SDRS):

    • Initial virtual disk placement (Pod – pool of data stores)
    • Out of space avoidance,
    • I/O load balancing,
    • Virtual disk affinity anti-affinity
    • New man agent object,
    • Storage equivalent of SRS clusters,
    • Consists of similar data stores,
    • storage load balancing domain.
    • Storage Policy based management (SPBM),
    • 2TB+ LUN support. vSphere 5.0 provides support for 2TB+ VMFS data stores.

    VMFS-5 – Scalability and performance improvements, increase limits of the file-system (limits a FS size that increases when the FS extends/grows), Reduce SCSI reservation with VAAI primitives. Plumbing of UNMAPs (Space reclamation on TP LUNs). Set the stage for future proofing, more efficient snapshot facility within VMFS, support for 2 TB plus files.

    vStorage API for storage awareness (VASA) – to allow array vendors to provide more info about the storage array, (drives, RAID level etc).

    VAAI – now doing NAS and Thin Provisioning (pass info down to array to clear up the space). Ability to warn when the array runs out of space.

    Storage vMotion Enhancement – speeded up, now snapshot VMs can be storage vMotioned.

    Storage vMotion snapshot support – Allows Storage vMotion of a virtual machine in snapshot mode with associated snapshots. You can better manage storage capacity and performance by leveraging flexibility of migrating a virtual machine along with its snapshots to a different data store.


    LLDP – Link Layer Discovery Protocol) – bit like CDP

    Netflow – inter-vm traffic, vm-physical infrastructure traffic – (vDS sends records to 3rd party collectors such as NetQoS and NetScout)

    DVMirror – can monitor traffic inter-vm or intra-vm.

    NIOC (network IO control)– network prioritisation (I/O shares) at VM level – QOS extended to network infrastructure – workload isolation between tenants. – done through vDS resource pools and shares within them.



    vMotion – support for up to four 10GBPS or sixteen 1gbps nics –single vMotion can now scale over multiple nics (load balance). It can now slowdown during page send (SDPS)(slows down the VM slightly to ensure the copy can complete) feature throttles busy VMs to reduce timeouts and improve success. vMotion can now ensure less that 1 second switchover time in almost all cases. It can support higher latency networks (up to 10MS), improved error reporting, resource pool integration by vMotion now puts VMs in the proper resource pool.

    DRS – integration with vShield, support for vShield agents (zones, edge, app, endpoint), a DRS/DPM cluster hosting vShield agent VMs, resource pool management consistent for clusters and non clusters being managed by VC. Resource pool settings stored in VC for non-clustered hosts (same as clustered hosts), enables support for stateless ESXi hosts running in standalone/non-clustered mode, prevents direct host access to resource pool settings managed by VC, improved behaviour when host disconnected from VC, if host loses access to VC users can connect directly to the host and override VC to take control, no longer requires restarting the VPXA.

    Cloud Level Scalability:

    • 1000’s of VDC’s
    • 50,000+ VMs
    • 2000 hosts
    • 150,000 Objects
    • 200+ concurrent administrators

    Can now run on Linux and also comes with a downloadable appliance.

    FDM (fault domain manager) – Replacement for VMware HA. no more AAM

    • More reliable
      • Deploys and reconfigures within seconds, regardless of cluster size.
      • Uses multiple channels for agent-to-agent communication: both network and storage
      • Removes dependencies on commonly misconfigured services (e.g., DNS)
    • New and Improved features
      • Management network partition support (new)
      • Single HA log file per host and syslog integration (new)
      • Host isolation response (improved)
      • Admission control (improved)
      • Agent error reporting (improved)

    Flex Based Client:

    • Empowering the Administrator
      • Centralize all VMware Administrative User Interfaces
      • Seamless integrate with all VMware solutions
      • Create a common user experience
      • Common “Look and Feel”
      • Single Sign On (SSO)
      • Scale to the Cloud
      • Support multiple platforms
      • Provide for ease of extensibility

    Its worth noting this is better than the web clients prior to vSphere 5.0 but still has a little way to go to match the vSphere Client.

    Update Manger – can now do delayed and staggered upgrades – can upgrade a Virtual Appliances.

    And for VDI:

    Accelerator. An accelerator has been delivered for specific use with View (VDI) workloads. With this option configured in ESXi, a read cache is constructed in memory that is optimised for recognising, handling, and de-duplicating VDI client images. The cache is managed from within the View Composer and delivers a significant reduction, as high as 90% by early estimates, in IOPS from each ESXi host to the storage platform holding client images. This reduction in IOPS enables large scaling of the number of clients in case multiple I/O storms, typical in large VDI deployments, occur.



    (I’ll keep updating as I find them):

    Post to Twitter

    Jun 242011

    First off credit for this post has to go to Brian Lyttle for his 4.1 USB blog post.

    Not to much has changed from the 4.1 methods of booting the install for vSphere 4.1.

    The method I used to follow was Ivo Beerans manual syslinux method. This works ok if you manage to follow every instruction perfectly. But in a attempt to make my life easier I found this method is much more straight forward and is easily repeatable.

    By using UNbootin to create the USB file, life become much simpler. It still creates a syslinux.cfg file but it does all the hard work for you.

    I’m a Windows desktop guy so UNbootin is great for my Windows 7 desktop but its also available for Linux and MAC OS.

    Here’s the auto configured syslinux file contents:

    default menu.c32
    prompt 0
    menu title UNetbootin
    timeout 100

    label unetbootindefault
    menu label Default
    kernel /ubnkern
    append initrd=/ubninit -c boot.cfg

    label ubnentry0
    menu label ESXi-5.0.0-381646-standard ^Installer
    kernel /MBOOT.C32
    append initrd=/ubninit -c boot.cfg

    label ubnentry1
    menu label ^Boot from local disk
    kernel /ubnkern
    append initrd=/ubninit

    Create the bootable USB Flash installation drive.

    Download UNbootin from sourceforge.

    Once you have stated UNbooin.exe select Ubuntu from the distribution dropdown menu (I left the version at the default).


    Select the ISO file you want to be able to boot to.

    If an overwrite messages appear just accept them.

    Dont click ‘reboot now’

    THATS IT… you now have a bootable USB Flash Drive with vSphere Hypervisor 5.0 instiller package.

    Ivo’s Kickstart files can be used to automate the instillation of the hypervisor should you wish to do so.


    To complete the instillation and configuration follow……………

    Post to Twitter

    Twitter links powered by Tweet This v1.8.3, a WordPress plugin for Twitter.