Now that 2018 is here it is an ideal time to clean up your estate. The problem is that rightsizing and cleaning up the estate isn’t always a straightforward proposition, especially when done by hand.
Correctly rightsizing a VM is a black art. You can look at a point in time or a week but that server may not peak until one or two specific points in a single month. It takes skill and time. The other big issue is that doing such profiling in a large estate by hand doesn’t stack up financially.
Out of the box most hypervisor platforms don’t really have rightsizing functionality / reporting built in. The story doesn’t stop there. Buying in a solution takes time and obviously costs money to run.
Recently I have written a lot about Embotics. As well as being a pretty good CMP (Cloud Management Platform) it installs really quickly. I setup a POC inside thirty minutes.
So what? It means that within a very short time the administrator can see what is oversized or conversely undersized and give suggestions on rightsizing. Obviously, the longer it is left to collect statistics the more accurate it will be. I’d suggest a month at least to get a good grounding.
Sure there are other solutions but I like an easy life. It is simple to install, use and will even do the resize for you if you so desire! Other systems can do this but they aren’t as quick to install and certainly aren’t as easy to use.
Check out the video I did and see just how easy it is, honest. Doing a rightsizing exercise is the way to go with any decent sized estate. It had been known for Embotics to be run in a brownfield site and recover enough resource to actually pay for the Embotics licences and support for the year.
Scripting can be scary and dangerous at scale ;). Worse, management have cottoned on to the fact that scripts can be dangerous. Result? Management take away the scripting tools. Yes, truly. I have seen it happen. Goodbye ad-hoc scripts. Help avoid them taking away your valuable scripting tools by reading my article!
So I was having lunch with a ex work colleague last week who now works for one of the major cloud providers and we started chatting about companies who just don’t get a very major cloud related point….
Cloud is not about servers, but services.
It may seem a somewhat obvious statement but there is still a huge element of “I want to own everything I use.” The way to look at cloud beyond a quick VM here or there is that most cloud providers already provide the services for your application requirements already…. Database, check. Web services, check. Service x, check. Have you ever seen how many services Amazon has? A lot.
The point is recreating everything from on-premises in IAAS, standing up your own virtual servers is overkill, kills the value proposition and misses the point.
Sure there are some specific cases (Security, compliance, contractual obligations and such) where an administrator will need to keep everything on “infrastructure” they control and manage. If that’s not the case, your doing it wrong and in a very expensive, long winded, re-inventing the wheel fashion.
Why stand up a full SQL server when you can get a DB service for a fraction of the cost and without any of the management overhead but the same (or better performance).
Lab in a box (nested lab) nerds such as myself love VMWare workstation. It’s the base upon which we build our labs as it supports ESXi out the box. I decided to take the latest version for a spin.
In my opinion Workstation 14 is a iterative release. There are a few new features that may be useful depending on your requirements. VMware Workstation 14 brings:
New CPU support (Ryzen etc)
A new hardware version (14)
Secure boot for VMs
virtual NVMe support
In terms of software there are several new items. Aside from the support for new operating systems (Ubuntu 17, RHEL 7.4 etc).
One interesting new feature is native OVF deployment. I learned that it is now possible to deploy the vCenter/PSC directly as a VM on the local workstation.
This makes things interesting because it removes the need to install ESXi and configure it out. Essentially popping out a PSC/vCenter is as easy as answering a couple of questions and half an hour later, a brand new vCenter! Makes life easier as your dont have to spool up an ESXi server.
Other features support for virtual TPM (Limited use case scenario in my opinion but you can use it as you would for Windows encryption and such.) and VBS support (Vitualization Based Security). VBS is tagged to become the next big thing according to those in the know regarding security. Support is dependent on the OS to be installed by the way!
Lastly, and quite interestingly, it looks like when installing systems there are new options! Virtual NVMe is supported. To quote from the VMware blog:
Virtual NVMe support Workstation 14 Pro introduces a new virtual NVMe storage controller for improved guest operating system performance on Host SSD drives and support for testing VMware vSAN. NVMe devices require virtual hardware version 13 / ESXi 6.5 compatibility and later.
The testing VMware vSAN certainly looks interesting.
All in all the upgrade looks to be worth it depending on your usage scenario. Personally I am quite looking forward to experimenting with the NVMe component above all. It was possible previously but required some kludges to make it work.
So I was fortunate enough to be given the opportunity to attend the NSX test drive experience. It was very worthwhile. So what’s it all about?
Firstly, Dell/EMC/VMware looked after us very well. I certainly wasn’t expecting bacon and egg sarnies in the morning ;). Anyhow… the course.
The class size was 16 (The maximum) and was delivered at EMC Manchester. Essentially it was quality as well as quantity 🙂
So it was described by several of the trainers as the ICM (Install, Configure & Manage) without the fluff. To be sure, it truly is a techie course for techies. No marketing slides to be seen (Mr VMware, please don’t take that as a hint 😉 )
The whole point of the course is to help people understand NSX and the advantages it brings and how NSX works under the covers. There are too many advantages to list in this short shout out but the way the course was delivered it absolutely made sense.
Day one was how VXLAN works, how to install and configure NSX and the theory. It was almost identical to those courses that you beg management to send you on.
Day two delved even deeper into the whole security setup and how the whole security functionality in NSX works. It may seem simple on the surface but there is a huge hidden layer of complexity if you want to dive that deep.
Other items included touching on Orchestration and how to setup NSX in a multi-vcenter environment.
These courses are free. If you have a good account manager just ask. I went from knowing a little about NSX to understanding the vast majority of how a basic NSX setup works.
I found it provided an excellent opportunity to the question of “Where do I even start” with NSX.
For anyone that is interested in a course 101 tour of setting up NSX I will be be writing an article that will be appearing on SearchVMware soon.
It would be remiss of me to not just say a big thanks to Mike, Phil and Kaela for making it happen. You guys rock!
Every IT admin has disaster stories of data loss. Most people have seen the picture of the IBM array that fell through the upper floor of the DC.
Unfortunately not every admin has good backups though. Admittedly, it is not something admins like to think about too much. Speaking to KrollOntrack revealed an interesting story about virtual data loss.
Whilst everything about virtualisation gets quicker, automated and more densely packed the risk of data loss increases. It only takes one bad action, intentional or otherwise and its a case of “Dude, where’s my VM farm gone.”
Speaking to Ernesto @ Ontrack it is a common occurrence. Reasons for data loss include many different scenarios… human error, power loss, malicious deletions, floods and natural disaster, physical array failure… The list goes on.
“We are not here to sell you anything” was a true reflection on the reactive service that you only need when something bad has already happened.
Hopefully I will be writing an interesting new in-depth article on this for those that are interested in the how, the where and they why but for now, if you want to hear some interesting stories I recommend you drop by their booth @VMworld.
And as a parting thought, to put it into context, shipping an entire array to the clean room for recovery is quite a normal occurrence! Feel you inner nerd.
The big message from this VMworld is as to be expected, hyper mobility at both the client end and the server/cloud end.
Any solution, any device anywhere is already very well known for end user computing but VMware have put their money where their mouth is and added support for Chrome devices for Workspace one. It now means that all the major players are supported.
This whole any device scenario gels with something I had noticed in general as I attend various events… more people than ever are utilising tablets and similar devices rather than classic laptops. Some misguided people even had the Ipad Pro *JK
Moving to the cloud side, it delivers HCX technology. The HCX technology allows the migration of workloads between on-premises, cloud or other other solution. We have all been there, where we have to do a migration and the downtime whilst a server is migrated. HCX solves that solution. It provides a way to do away with that downtime. It isn’t available yet, but it’s coming and it will make life very intesting. It is part of VMware’s play to be the glue that connects all the clouds together.
This change (which I assume includes parts of NSX) has vast ramifications. Think about it this way.. If you have one network that expands to cover all your environments where the VM sits is no longer important. It could be on-site, on cloud. No one cares (except the beancounters!)
This also has huge ramifications for DR. Normally DR, even virtual DR means that failing over requires some manual or automated configuration (ie IP addreses, host names, firewalls, applications etc.)
Reconfiguring that virtual machine, under pressure of a real DR could be a real pain. That pain is about to go away. With the new VMware stack it means that in a DR scenario bringing up the DR instance becomes a much simpler scenario. No having to re-ip or rename virtual machines and then hope for the best.
Last year we got teased about VMWare on AWS but now it is time to show about it. We all know the VMWare / AWS is the future. How to get on to it? Depends on your scenario:
Maintain and expand
Consolidate and migratie
vSphere can help with all those scenarios. Build your vSphere stack in less than 2.5 hours !
AWS is a full SDDC stack (vSphere, vSAN, NSX) deployed on bare metal and vCenter as control plane. It is that simple.
Hybrid link mode for single pane of glass, tick. Everything just works as expected as it is just vSphere under the hood. As I alluded to recently in an interview about AWS and vSPhere, it is VMware cloud. Support, VMware, Billing. VMware. It is VMware (just to push the point home.)
AWS do get a look in too. AWS serfices such as RDS, S3, IAM can also be consumed by VMware on AWS cloud customers.
vSphere on AWS Nodes consist of:
512 GB ram
14 TB NVME storage, 10 usable.
Other new interesting features include elastic scale. There are also several consumption models including consumption based, ad hoc (credit cards, PO’s etc and lastly, and perhaps most importantly, Hybrid Loyalty program.
Oh and lastly, VMware take care of your patching. More on this later as I am sure a lot of people are interested in that (for better or for worse).
After a bit of travel I have arrived. I am expecting many new and interesting things this year, not least AWS and VMware, all the items they teased us with last year should all be unleashed upon us.
For those returning from last year, the whole VMworld layout is completely different so it may take some getting used to! The new badges are useful though because the password for the wifi is on the back of the rather large attendee pass (as well as a map of all the locations!)
Don’t make the same mistake I made 😉
Need to find an AWS person to talk to and get some VMware on AWS exposure…