General

Rig for Silent Running, and some industry stuff

Hey all!

I apologize for the lack of post frequency at the moment – I’m in hard study mode for two certifications concurrently:

MCSA (So the 70-412 exam) and the Appsense Certified Professional exam.

We’ll certainly get to end user profile management and VDI layering at some point soon in this blog!

Speaking of VDI, Teradici released new firmware for zero client endpoints for both the Tera 1 and Tera 2 chipsets. This is a pretty important release if you’re running Horizon View 6 or playing with Amazon Workspace. TEST it in your environment before release though, as I’ve heard of performance problems if you haven’t upgraded to Horizon View 6 yet… and some connection problems if you load balance your connection servers via NLB or a hardware loadbalancer.

Link to the firmware: https://techsupport.teradici.com/ics/support/DLRedirect.asp?fileNum=1504&deptID=15164

.
One of my colleagues (And pretty swell guy overall) has had some fantastic results using EMC’s all-flash XtremIO Array and compression/deduplication- A linked clone pool of 1050 desktops was only using 690G of storage… a compression/dedup ratio of 5.1:1 … pretty nifty.
Of course, that’s just initial pool creation. I’m curious to see what the storage utilization looks like after it goes into production.

I’m building a physical server for my current client- and not one that runs a hypervisor. It feels weird. I’ve been such a pro virtualization guy for so long that the last server that I popped a Windows Server installation disk in was destined to be a SQL super computer maybe 24+ months ago. Oh well.

For the record, while the task is pretty specialized I’m pretty sure it could be virtualized. The limiting factor here is the server requires a pretty huge PCI-E card, and the client is running Cisco UCS blades that can’t handle it.

Speaking of UCS: I’m not more than entry-level skilled on the ways of Cisco UCS hardware – The team at Varrow have some UCS superstars that I rely on to get the hardware set up right. I’m going to lean on them a little bit as I learn because my current client is running 1 or 2 nics in each blade for ESXi (I would expect at least 6- 2 for management, 2 for vMotion and 2 for VM traffic) and no QoS. I want to try to take care of those issues before I leave or there will be some network bottlenecking as they grow to their intended scale. Infrastructure plan and designs are important!

Exam study blueprints – a change of pace

Hey all!

I received some pretty insightful feedback on my 70-412 exam prep blueprint series, as well as my frequency in posting for those studying for the MCSA/MCSE (Sorry again for the hiatus!).

Based on both the Extra-Large size of taking the certification study by section (3200+ words) being harder to study from, as well as Epic length posts taking much longer to lab out and write up on my end, I’m changing the format to 1 post per exam topic.

Thanks for the feedback guys!

You can reach me via email: peteDOTflutyATgmail.com

Via Twitter: @petefluty

Or the new Facebook page: https://www.facebook.com/EnterpriseEngineer

Checking In: Week of October 12th

Hey there Sportsfans!

Phew. I apologize for the hiatus if you’ve been following my Microsoft 70-412 blueprint series- The topics are getting intense and mentally draining which makes my desired 3 posts a week cadence pretty difficult. I could just send links to content, but I feel like that would cheat my readers a bit as I’m striving for original content.

The other hangup is technical – during the last series, I did bad things to my Lab AD Domain. Trying to leverage Dynamic Access Groups on top of it proved too much… I was getting all sorts of unexpected results.

Tonight I’m rebuilding my domain controllers, so hopefully I’ll be getting this show back on the road in the very near future.

Thanks for your emails and support!
Pete

From the Field: Be mindful of poorly configured VMs in a new environment

Hey everyone!

I just completed a multi-cluster upgrade to 5.5 U1b for a client, and found that one of the Virtual Center machines was running extraordinarily slow. A quick investigation showed the machine running at full capacity (sustained 100% memory(14G), 90-100% CPU(4vCPU)). This machine also required VMware Support’s help to manually recreate registry keys to remove vCenter 5.1, but that’ll be a topic for another blog post.

After some digging around and multiple sets of eyes, it was discovered that another internal Server administration group built all of the VMs for this clients Horizon View environment- They gave them the appropriate amount of memory, but set 4000MB limits on all of the machines…. and the site’s been running like this for 12-18 months.

The VDI guys just thought it was supposed to be that slow.

Lesson?

For anyone who works as a consultant in an unfamiliar client site, it’s important to check the basics when looking at a probable misconfigured VM environment:

– CPU Limits
– CPU Reservations
– Misconfigured resource pool trouble
– Excessive snapshot chains
– Host resource constraints
– AntiVirus issues
– Storage backend issues
– Host CPU Power Management
– VM free space and other guest OS issues

What are some things you would check for in a strange environment?

New Project: ESXi Home lab

Hey all!

I’m fresh off working with the folks at AppSense (http://www.appsense.com/) for a proof-of-concept at a hospital. I’m a road warrior cloud engineer, where technology changes constantly. Combine that with the fact that I have a new 1 year lab license from the AppSense engagement,  and partner licenses for other software… and I need to sit down and put together a real ESXi homelab.

We predominately work with Cisco UCS at work – needless to say, that’s beyond overkill for a lab and well beyond the budget in terms of acquisition cost, utility cost and acoustic sanity.

So what to do? I’m starting to sketch out requirements based on acquisition cost, recurring costs, and if my wife will let me bring it in the door. Namely:

  • Relatively low power consumption
  • Core i5 processor
  • Enough iOPS to run a VDI load
  • Sufficient memory
  • Un-noticeably quiet

With these requirements in mind I set off to Amazon. I haven’t purchased anything yet, but I’m leaning toward this configuration:

Reasoning:
– The Shuttle box is a nice form factor, able to run a Core i2/5/7 processor with 32G of RAM, has enough upgradeability for what I’m going to use it for, and a 300w power supply. I need to have at least two of them, so these things complement each other well.

-The Intel Core i5 is the sweet spot for price and performance- 3.0GHz with 4 physical cores – 8 with hyperthreading.

– 32G of RAM is the most this box will hold. Hopefully this will be good enough at this time for the lab. I foresee a time in the future where I’ll need to separate light and heavy workloads- a SQL server will run though that much memory all by itself, so I’ll have to be conscientious of that.

– 250G Sata III SSD drive- This will be the top tier storage in this system, the home for OS drives and VMware View replicas

– 3 TB Sata III 7200RPM drive – This is the bottom tier of storage- This is user and fileserver storage

– Intel QP NIC – The internet is rife with complaints of people trying to use the Shuttle box with ESXi and running into networking issues – the Intel QP nics are nearly universally supported by VMware and this card gives me sufficient numbers of ports for local, management, storage, Fault Tolerance and vMotion capability.

– SanDisk 16G low profile flash – I plan to boot ESXi 5.5 from this flash device, preserving the internal drives to handle the VM storage requirements instead of losing a portion to the hypervisor.

Now that I have all this hardware and assembled the machines, now what?
Click to move on to configuration!

Early Hiatus/ Employment Distinction

Hey readers!

It’s been a busy 2014 for me already, and I hope yours is also off to a great start!

I was honored with a highly coveted employment distinction this week to really start the new year right: my company has an MVP contest to recognize the individuals most responsible for our success as an organization, and all MVP winners are in the running for two MVP of the Year awards which include the much larger recognition and an all-expenses paid trip to Puerto Rico to party with senior management.

I won the quarterly distinction in Q2 for my efforts on a myriad of projects and business initiatives, and was selected to win the much larger MVP of 2013.

In addition to the normal holiday festivities that linger at our house (My in-laws will be leaving this weekend and my professional life should get back to its normal routine) today is my Birthday.

2014 is shaping up to be an exciting year!