I have had this book for a while and it’s the only Sybex book I have! So I thought it deserved a review.
The book is split up into the following Chapters:
- Performance Design
- Building Your Toolbox
- The Test Lab
This goes into detail about establishing good baselines for the various aspects of the environment. You can spurt out lots of figures, but if you don’t have a baseline that has been compiled over some time, how do you know if the current stats you are seeing are good or bad?
Also it goes into design considerations and mentions the impact of scaling up or scaling out, but also points out that you could ask this question to 10 different people and get totally different answers. When it comes to designing an environment there are many ways to go about it, and it all comes down to your requirements and constraints and everything in-between!
This goes into the key tools a vSphere Admin will need to have from capacity planning and bench marking tools. From using vCOPs, Log Insight to esxtop and its various modes. It also makes a good point about using the –a mode with esxtop. Then it goes into detail about vscsiStats and its various options. You even get details on tools you can use for stress testing, and how to go about using them.
This goes into the detail about building a test lab, why you should and the type you should build. There is no point building this super fancy lab if all you want to do is test some host upgrades. You can use VMware HOLs and AutoLab which are great, but having your own lab in my opinion is a good thing.
In my own lab that I have made at work, I have basically made clones of the prod environment including AD all on an isolated vSwitch with a virtual router. So the same IPs and logins can be used, and I can break and test it as much as I want without it impacting live and any issues I get in the test environment I know will happen in the live, so I can plan and adjust accordingly.
This is all about CPU, virtualization of it, the architecture of it, the CPU scheduler, CPU affinity, how NUMA plays a part, vSNMP, Hot Plug, Reservations and Limits. Then it goes into how to troubleshoot CPU performance issues from high ready time, high host CPU utilization and high guest CPU utilization.
This is all about memory, from TPS, ballooning, compression, swapping. Want to configure SSD caching? They have got you covered! You want some in-depth info on managing VM memory allocation…they’ve got you! You need to troubleshoot memory issues, then you have come to the right place!
It covers networking, from ideas about designing your physical network. Yes you are a VM Admin, but how the physical network is designed should be very important to you! It goes into detail about the VSS and the VDS. Also how selecting host hardware can impact the network design, and finally troubleshooting networking issues, and how to configure traffic shaping correctly
Is storage, it has a breakdown of the various types from FC, FCoE, iSCSI, NFS, VSAN. It then goes into detail about different RAID levels, the different types of multi-pathing and Storage I/O Controller. It even goes into datastore sizing, which has always been a mixed bag of opinions! The age old “500 Gb VMFS and no more than 15 VMs per datastore”, but as they say in the book “Numbers without context are meaningless”.
It then goes into some detail about Flash Read Cache which I thought was nice. There is also a section of the age old VMDK vs RDM debate and details on the various different types of VMDK types and finally troubleshooting storage issues.
Overall I think this book is brilliant, @ uses this book too and it should be part of every VMware admin/designers bookshelf.
I didn’t like the paper that was used in the book, it feels cheap! But lets be fair, if that is the only criticism I can give with regards to this book, then I have to say it is a must buy!
It is a brilliant companion for people doing the VCAP-DCA/DCD, there is sooooooooo much gold in it and lots of good reminders!