ANZ Pied Piper

In today's article I wanted to share some work I have been doing recently, specifically a prototype for an on-prem Storage-as-a-Service with DellEMC PowerStore A few years ago we used to hear a lot the "we have a cloud-first policy" and this of course meant "public cloud". The promise of lower cost pay-per use, self-service and automated IT consumption attracted many customers. Many of these organizations discovered the hard way that the financial aspect of it was not exactly "low cost". Public cloud makes solid sense for a bursty application but your average workload can typically run on-prem for up to 50% less However the self-service and automated IT consumption capabilities changed everyone's perspective of how IT resources should be consumed. Organizations nowadays expect that capability on-prem as well. If they succeed in getting it they will have the best of both worlds: empower the end users, operational efficiency and financial savings The...

PowerFlex (formerly ScaleIO and VxFlex) has long be one of the best choices for Kubernetes persistent storage. For a long time, ScaleIO was one of the handful of storage drivers available in-tree, even when CSI was not around. The thing is that PowerFlex is a scale-out software defined solution, much like Kubernetes itself. This makes it a great match. But also, in a world where we keep deploying scaled out high performance containerised applications such as ElasticSearch, MongoDB or Cassandra there is an ever increasing requirement for high IO low latency persistent layer. PowerFlex is a great choice for these workloads  With that in mind a colleague of mind has put together a series of videos showing how one can leverage modern DevOps tools like Ansible and Terraform to deploy such solutions easily and efficiently. Terraform is used here as a provisioning tool and Ansible for Configuration Management In the first video Jonathan uses Terraform to deploy a number of Virtual Ma...

In today's article I wanted to introduce a video I have been working with some of my colleagues from the IaC Avengers team. We had done some work with Service Now before, in terms of custom catalogs and even approval workflows. You can refer to this article to see the details What we had never done is to manage resources in more than one location. The Avengers team has a strong presence both in Australia and in Singapore, so we decided to create a solution such that from a single Service Now developer instance we could manage resources in both of our datacenters. As explained in previous articles this requires the developer instance to be able to reach out the internal infrastructure. This can be potentially a tough task due to firewalls and other security measures standing on the way. For that purpose Service Now provide the Mid server . This is a very clever piece of software that will run on either Linux of Windows and allow Service Now outgoing API calls to reach your on-...

Today I wanted to share a video that my colleague and fellow Avenger Masanori Nakamura has created. He is a presales engineer in Japan and is very knowledgeable with our VxFlex product. DellEMC VxFlex is a software defined storage solution that can be deployed either on a 2-tier or as a a hyperconverged architecture. It is typically consumed through appliances or as a fully integrated rack solution (formerly called VxRack Flex) However, since it is a software defined solution, it can be deployed on top of virtually any IaaS. In this case, Masanori-san deploys it on top of GCP.

Kubernetes can be hard work. The success of distributions like Openshift, Rancher or PKS is due to the promise of making it easier. Still a large part of the user base (in fact this still has the largest market share of any on-prem Kubernetes distro) choose to deploy plain vanilla Kubernetes to take advantage of the super fast-paced innovation However this is usually a path that leads to long hours trying to figure how to make things work and overcome many problems. This is painful even though the community out there is massive. With this in mind one of my colleagues ( Deepak Waghmare ) took upon himself the task of simplifying and streamlining some of these tasks as much as possible. He created an Ansible collection with several roles. This collection is now published in Galaxy The collection is primarily aiming at the the simplification of the installation of the CSI driver in DellEMC storage arrays. All these arrays follow the same pattern by using Helm, so with a littl...

The Ansible community is massive, so it is no surprise that version 2.9 came out with more than 3600 modules. This number keeps growing and it is motivating a change in how modules are distributed in the future. To read more about these changes you can read from Jeff Geerling himself However sometimes still you will come across with either: some functionality that hasn't been implemented on a module or a target for which there are no modules What do you do then? You have a few options: create your own module. This can be done in Python and even if you are not a Python guru you can find many tutorials that don't look intimidating at all you use the "shell" or "command" modules to run some other script if your target can be managed through a REST API then you can use the "uri" module REST API's are the basis for many automation tasks nowadays. An advantage of building automation through the API is that the workflow you build...

My colleague (and fellow Avenger) Andrew Vella has a strong developer background with experience on SaaS offerings, so when we came up with the idea of using Service Now to drive our Ansible playbooks he was very quick to put this hand up. This combination of Ansible and ServiceNow is an integration that we see many customers trying to build. Ansible's growth is phenomenal and is quickly sticking out above other Configuration Management competitors. ServiceNow is also dominating his market Andrew has promised to share his learnings in a blog when time allows. But in the meantime I can explain that he: created a developer instance in ServiceNow. This is free and you can keep it active as long as you don't forget to use it at least every 10 days installed the mid server inside the datacenter created a custom catalog with entries and an approval workflow configured REST API calls to trigger playbooks that are stored in AWX (the community version of Ansible Tower) mad...

In the previous article  I explained that the configuration management ecosystem and the automation in general is too crowded and dynamic and it is easy to make the wrong decision that can be costly in the long term. In my opinion, one tool is emerging as the winner above the rest, and that tool is Ansible. It is no surprise that more and more modules are being released to manage other aspects of the IT infrastructure including on-prem hardware. One example of that is PowerMax/VMAX which got its first release in June 2019. This first version covers all the basics of storage provisioning including snapshots https://github.com/dell/ansible-powermax It leverages the popular PyU4V python library  which wraps the REST API that lives in Unisphere. The library provides way more functionality than that, so it wouldn’t surprise anyone to see more versions in the near future that provide more functionality. The “Product Guide” document in the “docs” folder of the DellEMC repo prov...

In recent years, as cloud operating models have become more pervasive, organisations have started to adopt DevOps culture and processes to break previously existing organisational silos such as Development, Operations and QA. A key element for this transition is to treat Infrastructure as Code and to employ specialised automation tools. This is reporting benefits such as: Improved agility by eliminating manual or siloed tools which translates in significant reduction of provisioning times Minimise provisioning errors       Ensure consistency Ability to manage larger environments typically associated with scale-out microservices architectures Lower operational cost  The right choice of tool is essential to achieve these outcomes. Instead of leveraging element-specific scripting tools organisations need to move towards automation tools that can: address the configuration of large parts of the infrastructure so that multiple teams can “speak” the same la...