vROps 7.0 – Self Driving Operations (Part 01)

Ever since we hear this term Self-driven Car we develop this notion of a vehicle which has the built-in intelligence to take decision while moving on the road. There has always been approach in software industry to bring similar kind of intelligence in the data center where it is aware of the problems and is capable of self remediation. There are many tools available in market which claim to bring such intelligence but in my opinion there are very few tools which are actually capable of proving it. I have always been fond of VMware vROps and its capabilities and today I would like to share some of the facts which are offered OOTB with its latest release vROps 7.0 towards self driving data center approach.

pic01

There are basically three pillars which define Self Driven Data center :

(i) Continuous Performance Optimization

(ii) Efficient Capacity Management

(iii) Intelligent Remediation

If we look at the getting started page of vROps 7.0 all three pillars mentioned above are tightly incorporated within this landing page.  Let’s first understand how does Optimize Performance widget in vROps 7.0 is related to Continuous Performance Optimization

What’s Your Intent..? With vROps 6.7 there was an option to define Operational Intent for your cluster where customer can choose: (i) How densify customer wants to run their clusters. They can choose between Balanced or Consolidate option based upon which vROps will place workloads in that cluster. If they select Balanced then vROps will make sure that workload is equally distributed among all the clusters and no one cluster is over utilized compared to the other

(ii) Another option was to specify how much headroom one would like to keep in the cluster. So e.g. if my cluster has 10 cores and 100 GB of RAM and I select 10% as the cluster headroom then at any given point in time that cluster will not consume more than 9 cores and 90 GB RAM. This headroom capacity will be reserved for unexpected events or  satisfy future cluster demands.

However in vROps 7.0 VMware has one step further and introduced Business Intent. Now while defining operation-ability both of these factors (Operational + Business Intent) will play key role. Under business intent customer can define how do they want their workload to be placed within clusters? They can choose to do workload placement:

(i) Based upon the vSphere tags defined on VM or ESXi hosts (like Windows, Linux etc)

(ii) Licensing guidelines like Oracle or SQL vm should run on designated ESXi hosts

(iii)  ESXi/VM to be compliance for regulatory authority like HIPPA, FISM, ISO, PCI etc

(iv) Workload placement based upon SLA like business critical vm should run on Gold tier and Non-business critical vm should run on Silver/Bronze

Once we define it in vROps then it will keep monitoring the environment by itself and if there is any violation to this intent vROps will perform various DRS moves (within or across cluster) to bring the cluster in Optimized state

How to Turn On/Off Self-driving Operations: There are primarily three ways with which vROps can perform this optimization actions to your environment.

Run it now: Once the intents are specified the moment this option is selected vROps will take the control and identify wherever in the environment there are intent violations and based upon that it will prepare a list which all vm(s) to be migrated to which all cluster

Schedule: Since this is highly IO intensive operation and customer would like to perform it during non-business hours or over the weekend(s). So they can schedule this operation and vROps will perform this action on the designated day

Automate: Once this option is selected vROps will run the algorithm every 15 mins to check if there is any violation in the environment and once it identifies any violation it prepares those vm(s) to be moved to their respective host. Once the environment is optimized it goes into sleep for next 8 hours. Until then if there are any more violations then vROps will not report it and no further migration will happen. However this setting can be interrupted manually by Turning Automation Off and then On

pic04.png

Let’s try to understand from an example how does self driving optimization actually works within vROps

pic06pic05

 

Stage 01: Let’s assume we have two clusters named Cluster 1, Cluster 2 with each cluster having two ESXi hosts. We have defined defined vSphere tags to ESXi hosts as well as VM. Host 1 is tagged as Linux, Host 2 & Host 4 are tagged as Win & Host 3 is tagged as BSD.

In this setup I have defined my Operational Intent as Balanced (which means all clusters should run with equal amount of resource utilization and not one cluster is over-utilized compared to another one) I have also defined my Business Intent as VM with a particular OS (Win, Linux or BSD) should run on designated ESXi host.

Stage 02: Once I make this choice and handover it to vROps it will see the tag violations on Host 2 and Host 3 and migrate the workload running on it to the designated ESXi host to meet the business intent

However at this point my business intent is met but still Host 4 is over utilized as compared to Host 2 (Stage 02)

Stage 03: Since I had also defined my operational intent as Balanced so vROps will once again perform some DRS moves  and migrate some of WIN workloads from Host 4 to Host 2 — which is tagged to host windows virtual machine

 

How do I know which moves are done by vROps: This is a very valid question as how someone would know how many DRS moves vROps has made in order to balance out the environment. So the answer to that is Optimization History tab in vROps where you can view a time-lapse video of all the moves which vROps has performed to bring the environment to Optimized state. You can also get to see a graphical report for balanced threshold and DRS/Non-DRS moves performed by vROps in the past.

pic07

With all this intelligence customer can very much rely on this tool to help them to optimize their environment. So far we have only discussed about one of the three self-driving data center pillars, which I’ll be sharing in my upcoming blogs.

With this I end this blog here… I hope it was helpful for you.. till then Happy Reading 🙂

Advertisements

One thought on “vROps 7.0 – Self Driving Operations (Part 01)

Add yours

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Create a free website or blog at WordPress.com.

Up ↑

%d bloggers like this: