Tag Archive for ESXi

VCAP5-DCD Exam Review and Experience

VCAP5-DCD

So I successfully sat the VCAP5-DCD exam yesterday and thought I would jot down a few thoughts on the exam, how it went and some of the resources I used. The first thing I would say is… what a difference a year makes! When I sat the VCAP4-DCD in March last year, I was working as a Technical Consultant / Engineer, focusing on Wintel and Storage management / implementation. I spent over 2 months studying 3+ hours per night for the exam. Hard work, but absolutely doable and very rewarding, both personally and career-wise.

For the past 12 months I have been working as a Solution Architect for a Service Provider, primarily designing solutions for many medium sized businesses. The biggest part of my role involves talking to customers to define business requirements, translating these into logical and physical designs, and completing impact analysis to ensure all the components will fit together without any issues. Quite useful things to do regularly if you then want to take the VCAP! I can definitely say that I found the exam significantly less stressful the second time round, and my prep time was about a quarter of the VCAP4. My final week before the exam was spent as follows:

7 days before the exam At this point I was beginning to run out of time until the exam so had to be more selective in my review material. This included reading select sections from Duncan Epping and Frank Denneman’s excellent VMware vSphere Clustering Deepdive book and watching the last few vBrownbag VCAP-DCD courses from the inimitable Alastair Cooke & the APAC team.
Re-read my copious notes gathered over the course of my VCAP4 and VCAP5-DCD study.
Went through the blueprint again to ensure I hadn’t missed any key areas.

2 Days Before the Exam Read Paul McSharry’s Official VCAP-DCD Cert book over the course of a couple of nights. This is a great resource and I will be publishing a review on the book later this week, along with a wee Xmas give-away for one lucky person to win a signed copy of the book, so stay tuned! In hindsight I would have ideally given myself 3 days to read this and implement all of the end-of-chapter design exercises. The questions were invaluable practice for the exam and helped me build confidence in my knowledge immediately preceding the exam (never a bad thing if you’re finding it a bit daunting!). At time of writing I believe Paul is also working on publishing some more practice questions on his blog, over at www.elasticsky.co.uk.

Day Before the Exam Finished Paul’s Cert guide, then watched the excellent vBrownbag video by Harley Stagner. I particularly like this as it really goes into the thought processes in comparing different solution options and their subsequent impacts. If you have attended the official VMware design workshop, I also suggest this is the day to go through the case study again (I only had the VCAP4 workbook).
Completed the interactive exam simulation from the VMware MyLearn site – this is HIGHLY recommended as it will save you time in the exam not having to work out how to use the tool.

Day of the Exam I would describe the VCAP exams as very much a marathon, more than a sprint. I think its therefore prudent to approach this 4 hour brain mashing session in much the same way as an athlete might approach a 10k race. Have yourself a decent breakfast of slow-burn foods (porridge, Weetabix etc). Then shortly before you go into the exam perhaps have a banana. It will stave off the hunger later whilst giving your brain a bit of energy over a decent period of time.

Following a great wee tip from my fellow #LonVMUG member Craig Kilborn, I took a couple of ibuprofen about an hour before I went into the exam. They’re not exactly performance enhancing drugs (for those recommendations please see Lance Armstrong’s blog), but it wouldn’t be much fun to get a headache half way through the 4 hour session, especially if your exam centre wont allow water in the exam room! Water is important for your brain but don’t guzzle a load as you will end up having to spend the last 2 hours crossing your legs as I did in my VCAP4 (not fun!)…

The Exam Obviously I cant go into any real detail, but I will summarise some of the info already in the public domain, which I think is key to the experience. The exam itself was very similar to the VCAP4, though there are some slight tweaks, in that you now do fewer total questions (100 in all) and more visio-style design questions, a total of 6. If you are upgrading your existing VCAP4 I suggest concentrating some of your technical study specifically on the newer feature set as you will definitely be tested on these. For example, Jon Hall mentioned several times on one of the vBrownbag podcasts about Datastore Clusters.

You are now no longer able to mark or go back to any questions. I actually don’t think this is a great loss. The time required is so tight, you generally don’t have time to go back, and if you think you may have made a mistake then there’s no point dwelling on an answer as you cant fix it anyway! Move on and get more points elsewhere. I personally finished the VCAP5 with only about 5 minutes to spare, and the VCAP4 with one second to go! The tips I used for the exam approach in the VCAP4 still definitely hold true, with a few minor tweaks:

  1. Don’t pay too much attention to the clock, except when doing the Visio design questions, and maybe for the final 60 minutes or so. You know you will have 6 Visio questions so write 1-6 at the top of the page, and mark the start and end times of each Visio question. VMware recommends 15 mins per Visio question which is about right. Several of mine were under ten minutes and a couple of them took 16-18 minutes each. Don’t stress if one takes you a bit more time, but don’t let it go much beyond the 20 minute mark. At that point you’re better to move on and get more points elsewhere as you can still get lots of points for a “nearly right” answer. Fortunately VMware also seem to have made some improvements to the performance of the Visio tool, so it no longer seems to lag when you add a large number of objects. Based on 90 minutes worth of Visio questions, this will leave you with around 2hrs 15mins left to do the remaining 94 questions, or just under a minute and a half per question on average.
  2. Some of the Visio questions weren’t as clear in terms of language as I think they could have been. There were also one or two of the object types which didn’t quite make sense in the context. If you experience the same, again I suggest just do your best and move on. You can still get significant partial points as long as you have most of the diagram right.
  3. For each question, read the actual question before you read the case study information / description as it will help you to more quickly identify what information you are looking for and will reduce the likelihood of having to re-read anything.
  4. Once you have entered an answer don’t second guess yourself. Chances are your gut reaction is probably right as long as you have read the question and answers properly. I believe VMware have removed “trick you by missing or adding one key word in a sentence” type of questions, which is great. I don’t feel questions like this add any value, and certainly don’t prove your knowledge or lack of, one way or the other.
  5. I don’t believe there is negative marking employed, so any answer is better than leaving blanks. This is especially true if you reach the last 5 mins and still have some questions left. Speed read and answer quickly…

Final Thoughts I still very strongly believe that VMware should provide some kind of feedback in terms of the weighted scores against different subject areas or question types, much like you get in Microsoft or Citrix exams. If you were to fail the exam, it would give some good ideas as to the areas you need to study up on for your next attempt. Assuming you do pass, its still useful to know as it indicates areas you need to work on to improve your knowledge for your day job (after all we don’t just do these exams for the sake of it… do we?!).

Irrespective of your current role, I believe its definitely a certification worth going for. As an Engineer it helped me learn a lot more about design, business requirements, RTOs, RPOs etc, and that knowledge gained played no small part in me being selected for the role I am in today. As an Architect it has definitely been easier as some of the skills tested in the exam are part of my day to day role, but it reminded me of little bits and bobs I should be including in my design process and it’s one key step on the way to a possible VCDX attempt at a later stage.

Finally, I think this post is becoming rather wordy, so I will post my list of actual resources in a separate post later this week. If you are preparing for your VCAP5-DCD I wish you the best of luck!

Yet MORE Intel NUC Models on the way for your Nanolab!

For those of you who are regular followers of my blog, you will know I am a great proponent of the Intel NUC range for their low noise, low power, low(ish) cost, high performance and most importantly high WAF (Wife Acceptance Factor) features!

Unbelievably having only just announced their second generation triumvirate of models just 2 months ago (and due out in a couple of weeks), they’re at it again, announcing a third generation already! The new models include a pair of Haswell-based “Wilson Canyon” Core i3 / Core i5 processor options, featuring up to 4 USB 3.0 ports and a full size SATA connector and are expected to land some time around Q3 this year.

I have updated the CPU table with the currently available info on the new models, and will add CPU benchmarks once available on www.cpubenchmark.net (for consistency). This also includes the recently leaked specs for the new Gen 8 HP Microservers based on Intel Pentium / Celeron processors.

GenModelCores / Threads / Logical CPUsClock Speed / Turbo (GHz)CacheMax TDP (Watts)CPU BenchFeatures
1Intel Celeron 8472/1/21.1 / None2 MB17986None
1Intel Core i3-3217U2 / 2 / 41.80 / None3 MB172272None
2Intel Core i5-3427U2 / 2 / 41.80 / 2.803 MB173611vPro & VT-d
2Intel Core i7-3537U2 / 2 / 42.00 / 3.104 MB173766VT-d
3Intel Core i3-4010U2 / 2 / 41.70 / None3 MB152253VT-d
3Intel Core i5-4250U2 / 2 / 41.30 / 2.63 MB153572VT-d
1 (G7)AMD Athlon II Neo N36L2 / 1 / 21.30 / None2 MB12751None
2 (G7)AMD Turion II Neo N40L2 / 2 / 41.50 / None2 MB15946None
3 (G7)AMD Turion II Neo N54L2 / 2 / 42.20 / None2 MB251314None
4 (G8)Intel Celeron G530T2 / 2 / 42.00 / None2 MB351604iLO
4 (G8)Intel Pentium G630T2 / 1 / 22.30 / None3 MB352154iLO

IMHO you cant beat the NUC for its price / performance / noise features mentioned above. In an ideal world I would be happy to give up 2-3cm of extra board size to get some extra RAM slots and a second gig port on the VMware HCL in there, but as a tidy home lab solution they’re hard to beat!

As regards this latest batch of models, I personally still think the sweet spot is with the Intel Core i5-3427U DC53427HYE 2nd Gen model, which includes vPro for remote access, and will turbo to a handsome 2.8GHz for as little as ~£235 when I last checked. More than enough for most home lab requirements!

Maximising Perceived Memory Utilisation in vSphere

One of the few drawbacks of the Intel NUC, HP Microserver or most other small form factor motherboards is that most of these systems have only two memory DIMM slots. With current technology, this leaves you with a maximum of 2x8GB DIMMs or 16GB RAM in total. A couple of months ago the following tweet peaked my interest in trying to maximise the memory availability in my Nanolab environment, which consists of 2 NUC boxes with a total of 32GB of RAM in the cluster.

 

One of the ways in which VMware maximise VM consolidation ratios when virtualising is through a technology called Transparent Page Sharing (or TPS for short).

I was only running a relatively small number of VMs on my environment but was starting to run out of RAM and was disappointed with the levels of memory sharing I was seeing. What I had completely forgotten was that large memory page support is enabled by default, and these large pages will only be split down into 4k chunks in the event that my cluster was in memory contention.

I decided that as I am not running any particular high performance apps in my lab, I would prefer the visibility of how much RAM I actually still had available, instead of gaining maximum application performance through large page support. Enabling this was very simple, and simply required changing the advanced setting Mem.AllocGuestLargePage to 0 and waiting for TPS to kick in later that day.

My memory utilisation before was as follows:

Large Page Support Enabled

After it was as follows (you can see that not only can I see the RAM savings, but I have also added several more VMs in between screenshots):

Large Page Support Disabled

Assuming you have a large number of similar VMs within your home lab, disabling large memory page support can allow you to gain easy visibility of your maximum memory savings and actual available RAM. Implementing this in your production environments may not be ideal based on your specific workloads, however if your production policy is to be reasonably aggresive with memory overcommitment, I recommend you highlight this issue to your capacity management team to ensure they don’t go out buying extra servers or RAM unnecessarily early!

Further reading:

NanoLab – Harder, Better, Faster, Stronger Intel NUC Models Out Soon!

This story actually broke about a week ago, but its been quite a busy one for me so I didn’t get around to posting (other than on Twitter for those who follow me). I thought for people who may have missed it, it would be worth a short post.

In essence, for people who have held out from buying either an Intel NUC or even an HP Microserver for your home lab due to the lack of CPU grunt, good news is on the way! The specs were leaked last week for the new range of Intel NUC boxes due out in Q2, featuring Intel Core i5 and i7 processors. The specs were published by Computer Base and are as follows:

D53427RK - Rend Lake

D53427RK – Rend Lake

D53427HYE - Horse Canyon

D53427HYE – Horse Canyon

D73537KK - Skull Canyon

D73537KK – Skull Canyon

Looking at the new models the best (and most feature rich) for me is the i5-3427U D53427HYE (Horse Canyon – includes enclosure). This model includes vPro / AMT support, a superbly useful feature if you plan to run these machines headless, as I currently do. It seems strange not to include this feature with the i7 version (Skull Canyon – DC73537SY). The i5 is likely to be a little easier on the pocket than the i7 whilst still allowing turbo to 2.8GHz, and with a basic clock speed of 1.8GHz it will hopefully run a little cooler than the i7 (even the i3 chassis can get very warm indeed!). Both models come with USB 3.0 which is unfortunately of limited use, unless you plan to mount a USB drive into your VMs via VT-d which is now also included with both new CPU models.

Comparing the CPUs via their CPU Benchmark scores, we can see that the i5 gives a great performance leap from the older i3 line (DC3217IYE), but not such a great jump to the i7, which also doesn’t include vPro. I have included the scores for the HP Microserver line for comparison:

ModelCores / Threads / Logical CPUsClock Speed / TurboCacheMax TDPCPU Benchmark
Intel Core i3-3217U2 / 2 / 41.80 GHz / None3 MB17 Watts2272
Intel Core i5-3427U2 / 2 / 41.80 GHz / 2.80 GHz3 MB17 Watts3611
Intel Core i7-3537U2 / 2 / 42.00 GHz / 3.10 GHz4 MB17 Watts3766
AMD Athlon II Neo N36L2 / 1 / 21.30 GHz / None2 MB12 Watts751
AMD Turion II Neo N40L2 / 2 / 41.50 GHz / None2 MB15 Watts946
AMD Turion II Neo N54L2 / 2 / 42.20 GHz / None2 MB25 Watts1314

My guess is that two things will probably happen when it comes to pricing. The current line of NUCs will drop their prices a bit, and the new line will probably come in at a higher price bracket. This means a premium for people wanting the extra grunt, but better prices for everyone else! Personally I have not found any issues with the grunt I get from the 1.8GHz i3, especially when running off SSDs (where your bottleneck usually lies in a lab or production!) so I will probably stick with my i3 pair for now… at least until the i5 range become so cheap I feel compelled to buy a couple!

If I hadn’t already invested, I would be sorely tempted to start my Intel NUC lab with the i5 range, but if a key decision driver is cost, the i3 won’t let you down! 🙂

Other NanoLab articles may be found here:
NanoLab Articles

%d bloggers like this: