VMware EUC Conference 27-11-2012

Today I presented at the VMware EUC conference here in Perth on behalf of a client - it was a great opportunity to meet other members of the VMware community (both customers and partners) and discuss the success of our clients VDI deployment. 

If you have any questions about this particular VDI deployment or have general VDI deployment questions, please get in touch. 

To see a vendor case study on this deployment click here:

http://www.hds.com/assets/pdf/hitachi-success-story-hbf-uxc.pdf

To see a news article on this deployment, click here: 

http://www.itnews.com.au/News/293545,hbf-upgrades-it-infrastructure-for-virtual-desktop-rollout.aspx

#vForum2012 Part 1 of 3

As mentioned in a previous port, I recently attended the annual VMware Technology conference in Sydney #vForum2012.

As a follow-on from VMWorld in Barcelona, this two day event showcased not only the latest VMware products but also a number of partner products including storage, networking, backup, server and most importantly everything that is "CLOUD Computing".

While at the event, I attended a number of technical and networking sessions which provided some great insights and contacts. (During the sessions I took notes which I will eventually translate in to blog posts for each relevant technology).

The key themes throughout the sessions were:

End-User-Computing (EUC) & Mobility 

Simplification of desktop computing by delivering desktops & applications as a service which can be accessed via any device.
This is achieved using mechanisms such as VDI (Virtual Desktop Infrastructure) / “VMware Mirage” desktop management and application catalogs which are deployed using “VMware Horizon Application Manager”. As a side note - Applications within these catalogs can also be developed using “vFabric” (VMware’s highly integrated and flexible Spring Java development environment).
Another area of interest for me was a project (Horizon Mobile) that allows users to run a "virtual phone". This equates to having one handset that can switch between a work and personal phone at the tap of a screen - a great concept that almosts works as a phone hypervisor!

Cloud & 'Software-Defined-Datacenters'

The word on everyone lips is definitely "Cloud" - The use of this computing model is rapidly maturing with more and more companies now deploying cloud solutions through public: service provider datacenter, private: internal datacenter and hybrid: service provider and internal datacenter.
With cloud computing becoming a highly 'consumerised' product now comes the challenge of managing its consumption - "The IT Dept." is no longer seen as a barrier by the business.
VMware are leading the game in this space by providing highly simplified and integrated technologies – with the right configurations, organisations can now “drag & drop” application workloads into a public cloud environment should extra resources be needed (Strict SLA’s are of course required for this).

More content from #vForum2012 will be posted soon – watch this space!

Dismantling HDP Pools on Hitachi VSP

When Dismantling HDP pools, ensure you either go through or note the following steps:

1. Ensure Virtual VOL's aren't being accessed (check paths)

2. Remove Pool VOL'S from Pool(s). NB: The final Pool Vol will have to removed when destroying the pool itself as this holds the Meta data for the pool.

3. After deleting the pool you will notice  that 1 x Pool VOL will be 'blocked. or possibly more if you haven't removed them from the pool before deleting the pool itself.

4. To render these Pool VOL's useful again you can either "Restore LDEVS" (this can fail sometimes and if so you will be required to re-format the Pool VOL's. NB: LDEV Addressing will stay in tact)

5. Following this you can either remove the pool VOL's or reassign to another pool provided the RAID type is OK e.g. 'new pool' is built using the same raid configuration

vSphere Hosts - Max 'Open Virtual Disk Capacity'

If you are run "GIANT" VM's within your virtualization environment,
this article by @jasonboche is a great read - I began writing my own
but this is an excellent article - Thanks Jason!


http://www.boche.net/blog/index.php/2012/09/12/monster-vms-esxi-heap-size-trouble-in-storage-paradise/
 
In Summary:
 
vSphere hosts have a statically assigned "heap size" that
creates a theoretical maximum of 8TB of 'open virtual disk capacity' per host.
 
This is easily changable by heading to Advanced Settings (Softwre) under Host configuration in vCenter (Range = 16MB-256MB which can push 'open virtual disk capacity' up to 25TB).

#vForum2012 - Sydney (Nov 14th/15th 2012)

I have just returned from #vForum2012 and what a great event it was - the content, presenters, location and networking opportunities were excellent!

In the coming days I will be publishing some highlights from the event and also sharing with you thoughts on current and future VMware related technologies/partners.

With the increasing adoption of VDI it was great to see the new deployment approaches, especially around:

• Storage e.g. Nimble (Great to catchup with and meet these guys!)
• VDI-In-A-Box Solutions e.g. Pivot3
• Host side PCIe cache boards e.g. FusionIO

vSphere Host vs. VM CPU Mapping

From time to time I still get asked about the mapping between CPU's on the ESXi host and Virtual Machines it hosts (this usually stems from discussion around application licensing - more recently around Tibco) . Here is a good definition on how this works, which will hopefully provide some clarification:

If you do not have Hyper-Threading (HT) enabled it (ESXi) maps the vCPU to a logical processor which is a core, if you have it enabled it will map to a logical processor which comes from HT. But HT itself is not equal to real core, it is just a kind of an additional scheduler instance so the OS (ESXi) scheduler can drop its work there and continue without waiting for the work to be scheduled to the real core. E.g. if memory access is requested, the OS has to wait for that to be finished by the real core if HT is not available. With HT it can continue to schedule new jobs to the next instance without waiting. 

if you have any suggestions or adjustments for this, please feel free to comment!

VCP510 - Certified!

Today I sat and passed my VCP510 Certification Exam

Brocade Fibre Channel Upgrade

Below is a Fibre Channel upgrade (4GB to 8GB) I recently completed: Silkworm 4XXX's to DCX-4S's

This migration occurred during business hours without any interruption to business services (4 x SAN Arrays, 21 vSphere Hosts + variety of other open systems hosts).

Scope of the work included:

• ISL Trunking to ease migration bandwidth requirements
• Fabric migrations to FABOS V6.4 (V7 to come)
• Migration of IBM BladeCentre subordinate FC modules to access gateways using NPIV
• Decommission of old hardware

How Does Wide Striping / Disk Pooling Work?