TheSaffaGeek

My ramblings about all things technical


3 Comments

VMware vSphere: Manage for Performance Course Experience

Last week I was fortunate enough to be able to attend the VMware vSphere: Manage For Performance course. I did the lab related to this course at VMworld Europe last year and in my now increasing preparations for the VCAP-DCA exam I knew I needed to strengthen my troubleshooting skills and more importantly fine tuning my ESXTOP/RESXTOP skills. Quite a few people commented that they really liked my VMware vSphere: Manage and Design for Security Course Experience posting so I thought i would try do the same for this one for anyone interested or thinking about booking the course.

  1. Day one covered the first three and a half modules Course Introduction, Performance in a Virtualized Environment, Virtual Machine Monitor and part of CPU Performance. It introduced you to all the monitoring tools you can use with an introduction to the performance graphs in Virtual Centre and ESXTOP,how to utilise these tools to work out possible problems and what to look for that may be good indicators of problems. Probably the thing I enjoyed the most about this course was that there was loads of labs for you to learn how to do it all yourself rather than learning it off a PowerPoint sheet or your course guide. If you are a regular reader of this blog then you’ll know I’ve been trying to perfect my knowledge of ESXTOP/RESXTOP and wrote a blog about it too “Understanding and using ESXTOP/RESXTOP”. This learning of it was a great stepping stone for the skills they covered in the course and for a few parts the links and resources in my blog gave me an even deeper knowledge of ESXTOP. For the virtual machine monitor module it covered Software and hardware virtualisation techniques which i knew fairly well from my studying for my VCP exams and the labs for it were really great in covering how the usage of these varying techniques can really help with the performance of your virtual machines/environment. Next we got into a bit of the CPU Performance module which introduced the CPU scheduler, CPU Cache contention and the NUMA. As with the hardware/software virtualisation techniques i had a good bit of knowledge about the CPU scheduler and NUMA from my VCP studies but it was a great refresher on the NUMA particularly and allowed me to better understand how it works and how the misallocation of resources can impact your virtual machines due to NUMA. Frank Denneman has done two brilliant postings all about the sizing of VM’s and NUMA Nodes and ESX 4.1 NUMA Scheduling which covers pretty much everything you need to know about this feature and how to use it correctly in your environment.
  2. Day 2 we finished off CPU Performance by learning how to use ESXTOP and the performance metrics in vCentre to find and recognise possible cpu problems and how to fix them. Next we covered Memory Performance which was fairly straight forward in my opinion but did give great recommendations on how to utilise your memory effectively and how ballooning and memory swapping works and what the increase of these values means to the performance of your environment. Yet again Frank Denneman has covered these topics brilliantly in two blog postings Memory reclamation, when and how? and Disable ballooning? which I’ll personally be rereading through myself so as to better my understanding of how it can help/impact my virtual machines. Next we did the Network Performance module which covered all the varying network card options you can select,what each allows you to do,what additional features each one gives and how these features work. This was also a refresher for me due to my VCP studies but it did seem to alert a lot of the people on the course with me to the benefits of upgrading all your virtual machines to hardware version 7 and changing their network cards to VMXNET3. VMware have a great KB article on this Choosing a network adapter for your virtual machine. For the rest of the module it was yet again teaching you how to find and troubleshoot possible network problems using the performance charts and ESXTOP.
  3. Day 3 finished off the last three modules Storage Performance,Virtual Machine Performance and Application Performance. Storage performance was good and was very interesting to hear how many people don’t use thin provisioning due to their belief that it impacts performance in certain ways. I’m not going to get into it on here and I agree it does in certain instances but like I said to the people on the course with me I would recommend reading  the VMware white paper on it first and make your own decisions from there. There are also loads of top blog postings on the subject so I would also recommend reading a few of those (Duncan Epping’s and Eric Gray’s in particular). For the last two modules of Virtual Machine Performance and Application performance these were essentially just applying what you learn for cpu,memory and network to your virtual machines and what to consider for the virtualising of differing applications.

Funnily enough whilst on the course the latest release of the vSphere performance troubleshooting guide for 4.1 came out which is perfect post course reading material for me. Duncan Epping’s posting alerted me to the release so only right to point to his posting here.

Well that’s a high level review of what I learnt/was covered in the course. As with any course though what you get out of the course is very dependant on your knowledge of the product/s and even though I have a fairly good amount of knowledge on the product and features I did still learn a fair amount and it was a really great refresher on certain features in preparation for my VCAP: DCA exam sitting.

Gregg

Advertisements


Leave a comment

All Things Virtual 19

Since my last All Things Virtual there has been a massive amount of news coming out of the virtualisation arena. One of the main sources of these were due to VMworld Europe happening from the 11th to the 14th of October. I was fortunate enough to attend and I was very pleased with the amount of views my All Things VMworld Europe Day1, Day 2 and Day 3 blog postings received.

The week prior to me attending VMworld I was able to attend a day arranged by VMware and Alan Renouf where some of the top VMware GURUs came to London and gave some exclusive presentations on VMware API’s, performance, PowerCLI and Onyx. The day was seriously in depth and really opened my eyes to the amount of data and things you can get and do with the VMware API’s and ESXTOP. Alan has posted the slide decks from each session ran on the day on his blog.The sessions are listed below with their links, these are something I am planning to get through in my preparations for the VCAP exams

Due to there being such a large amount of time between All Things Virtual 18 and All Things Virtual 19 (a month and a half) there are a few articles on my list that are a few weeks old but nevertheless still brilliant and so I kept them on here.

As I have said before I ,like a large number of other professionals in the virtualisation field am preparing and studying for the VCAP exams. For my own usage and because amazingly a few people sent me messages saying that my VCP4 Study Resources(Part 1, Part 2) proved very helpful with their preparations for their VCP4 exams and were asking if I planned to do the same for the VCAP exams. So I have created a dedicated page for all the resources I will be using in my preparations and will constantly try to update it with any new resources I find. The page has only been up for a month and a half and already over 1000 people have viewed it which is really great and hopefully it is helping people find out about great podcasts/blog postings/videos that will help them be prepared for the exams.

Duncan Epping of Yellow bricks fame posted a very small but very important posting all about Storage IO Control Best Practices. If you attended either of the VMworld conferences then I would recommend watching the session Duncan speaks of in his posting “Tech Preview: Storage DRS” and another session i watched all about SIOC “prioritizing Storage Resource Allocation in ESX Virtual Environments using Storage I/O Control”. Speaking of SIOC Luc Dekens has posted a brilliant posting on how to automate SIOC.

While at VMworld Europe Mike Laverick of RTFM.com fame went around and got himself a bag full of Swag and is raffling it all off in aid of UNICEF. The competition is open to anyone worldwide so go get yourself a ticket and donate to a very worthy cause at the same time.

Eric Sloof posted all about the availability of the the Technical White paper for Application HA. My knowledge of Application HA has only started growing very recently after attending VMworld this year and actually seeing some of providers stands and exhibitions of the technologies and their capabilities. I haven’t as yet read the white paper but like so many of these great resources it is on my technical reading list. Eric has also posted a video of the installation of Symantec Application HA.

Simon Seagrave of TechHead fame posted about some great free NTP Time Sources and the NTP Pool project. I’ve already been using the same source for my NTP synchronisation for my home lab and it works a charm Smile

During VMworld Europe David Davis of Trainsignal’s VMware vSphere Video Training fame was able to do a video interview with VCDX001 Mr John Arrasjid. The video is a brilliant one for anyone looking to take on the VCDX certification as John is one of the main people involved in the VCDX and gives great insight into the certification and how it is all done.Jason Boche was also interviewed by David at VMworld US all about the VCDX certification here. These are two really great resources and have been added to my VCDX study list .

Speaking of Jason Boche he posted a brilliant posting all about  the conversion of CPU ready to %RDY in ESXTOP. I am currently learning and strengthening my knowledge on ESXTOP and how to utilise it for my daily job and Jason’s posting is a great resource for learning all about how the %RDY value is calculated. I would recommend reading through a posting I did a while back about Understanding and using ESXTOP/RESXTOP as this will give a great base knowledge before going through Jason’s posting.

Steve Jin of DoubleCloud.org Fame posted all about using vSphere APIs to Collect vCentre and ESX Logs.After having attended the VMware vSphere: Manage and Design for Security Course last week I’ve gained an even keener interest in learning all about collecting logs for security reasons as well as troubleshooting reasons and Steve’s posting was right on time for my interest in learning more about it all. I would recommend everyone learns how to do this as it will be a life saver at some point for you.

Vladan Seget posted all about a whitepaper by VMware on the top ten most forgotten things when building your disaster recovery plan. The white paper is a really great resource and has some very good pointers and recommendations when you are building out your disaster recovery plan. If you have  a disaster recovery plan or are planning one then I would highly recommend reading the white paper.

The next is an unusual one but one that I experienced myself and therefore feel I should link to it here even though I have already added it to my Upgrading from ESX,VMware Update Manager and Virtual Centre 4.0 to Version 4.1 posting. VMware have put out a KB article on an error where after upgrading your Virtual Centre Server to version 4.1 the transaction log for the vCentre Server database grows excessively large. Thankfully someone posted the KB article on twitter and i noticed it and fixed the problem before it severely impacted my Virtual Centre server.

Bas Raayman posted some really nice installation tips for installing VMware vCloud Director. I am yet to install vCloud Director to my home lab and so the installation tips are now part of my build documentation to go through when i finally get round to installing it.

Duncan Epping of Yellow-Bricks.com fame posted one of the best articles of the year over a week ago now all about VMotion, the story and confessions. The post covers the initial idea and building of a VMotion(misspelling intended) prototype to a number of top bloggers confessions on how amazed they were when they first saw or heard about VMotion and how it has impacted their work life and for some their career paths.

Well that’s all the postings of the past few weeks that have gained my interest.

Gregg


3 Comments

Understanding and using ESXTOP/RESXTOP

Lately I’ve noticed more and more people referring to stats and figures they have collected on their environments via ESXTOP/RESXTOP. I learnt about esxtop for my vcp exams but I have to be honest I haven’t really used it in my environment and recently I felt that I was surely denying myself and thereby the possible performance advantages for my environments by not using this tool to try find and fix any performance problems my vm’s/hosts may be experiencing.

I thought I would try write up a blog posting detailing everything  I had learnt but as is normally the case  some of the top bloggers out there have already beaten me to it and done such amazing jobs of it there’s no real reason for me to to do it. So instead I have collated all the information I have used to strengthen my knowledge on this tool and how it can help every virtual infrastructure administrator and or user to get the most out of their virtual machines and servers, especially with the imminent removal of the COS.

First is the ultimate resource for reference and learning what is possible with esxtop(I’m still amazed how much this tool can do and with update 2 there’s even more functionality). This VMware communities document has everything you could ever want to learn and know for esxtop and is staggeringly detailed whilst still being updated to include new functionalities and options. I’ve only managed to absorb some of the information off this document so far as it’s so much to learn and take in.

Next is a brilliant posting by Duncan Epping of Yellow Bricks fame all about the usage of ESXTOP and how he has utilised it with perfmon to retrieve all the desired data he needs for troubleshooting problems.

Jason Boche has also done a posting all about the capabilities of drilling down through the performance metrics to get very specific results.

Forbes Guthrie has done a brilliant vReference card for all the performance metrics you will see in your performance troubleshooting and monitoring. A very helpful “print out” capable card for quick reference as the same implies 🙂

Simon Long of The SLOG fame has done a very interesting posting all about using ESXTOP with VMware ESXi and is a very interesting look and example of how to do this monitoring once the COS is gone and ESXi is the only dominant host type. This kind of example is one of the main reasons I knew I needed to brush up my ESXTOP skills.

Kedrick Coleman pointed me to a very nice tool by VMware labs called ESXPlot which “is a GUI based tool that lets you explore the data collected by esxtop in batch mode”. I am yet to test this out but it looks like a very helpful tool in getting your performance data into once easy to monitor and even better to show to the boss/upper management the performance statistics of your environments/specific machines

Lastly if you were fortunate enough to attend or have a subsciption to the VMworld 2010 sessions then I would HIGHLY recommend the Troubleshooting using ESXTOP for Advanced Users session

Hopefully some of the resources help people strengthen their knowledge of ESXTOP/RESXTOP.

Gregg Robertson