OpenStack Cloud Application Development Quotes

“If data is stored in an environment that replicates data, the application should take care not to do its own data replication”
― John Belamaric, OpenStack Cloud Application Development

“The way you achieve “cheaper” is to make those standards open and free”
― John Belamaric, OpenStack Cloud Application Development

“Organizations may operate private clouds for a number of reasons. The cost of a private cloud, if well run, may be less than utilizing the public clouds”
― John Belamaric, OpenStack Cloud Application Development

“Put succinctly, IaaS provides the tools to “build” your systems from the ground up. PaaS allows you to “deploy” your applications, without needing to worry about the underlying infrastructure. SaaS allows you to “buy” your applications—you do not even need to deploy or manage them at all. This is a steady progression of decreasing control and complexity, while increasing direct business value”
― John Belamaric, OpenStack Cloud Application Development

“Elasticity adds yet another layer to an already complicated list of technologies and takes time and energy to implement and perfect”
― John Belamaric, OpenStack Cloud Application Development

“Something you won’t hear much of is that not all applications need to scale. Not scaling isn’t interesting. Not scaling isn’t cool. Not scaling won’t win you any awards. However, if you can avoid scaling, then you can focus your efforts elsewhere and greatly simplify your deployment. Some examples of applications or environments that may not need to scale”
― John Belamaric, OpenStack Cloud Application Development

“As a free solution, Sensu is probably a great place to start if you need to push beyond what will be demonstrated here”
― John Belamaric, OpenStack Cloud Application Development

“Generally speaking, it’s hard to recommend any of these as great solutions. Monasca should be a slam dunk, but it’s rather enigmatic and Ceilometer doesn’t provide a lot of flexibility when compared to systems like Sensu and Nagios”
― John Belamaric, OpenStack Cloud Application Development

“Ceilometer can measure things like load, and trigger alerts when certain thresholds are met. Unlike Monasca, it can also report detailed information on things like how much processor time was used by a given virtual machine or project”
― John Belamaric, OpenStack Cloud Application Development

“Monasca is a PAAS component that offers monitoring as a service”
― John Belamaric, OpenStack Cloud Application Development

“Lastly, there are new companies, such as Rancher Labs and projects like Kubernetes and Mesos, that are starting to provide container-focused orchestration solutions that live on top of, or work with, OpenStack. These are the bleeding edge of virtualization technology and as such are likely to see huge changes before mainstream adoption”
― John Belamaric, OpenStack Cloud Application Development

“Unfortunately, as with a lot of revolutionary technology, Magnum isn’t always available, and supported orchestration engines like swarm, Mesos, and Kubernetes could still use some time to mature and stabilize”
― John Belamaric, OpenStack Cloud Application Development

“In theory, containers on bare metal provide the best of all worlds. You get the efficiency /density of containers, the native management of VMs, and (according to some early benchmarks) nearly bare metal performance.”
― John Belamaric, OpenStack Cloud Application Development

“This could theoretically be achieved by provisioning a bare metal server, loading on a standard Linux distribution (or a more dedicated one for containers such as CoreOS or RancherOS), and running Docker or another container system on top of that. This idea has some advantages over bare metal alone, like being able to subdivide hardware. Unfortunately, it still lacks the orchestration and management capabilities provided by OpenStack.”
― John Belamaric, OpenStack Cloud Application Development

“The biggest hurdle with containers, however, is that they aren’t yet available as first order objects within OpenStack”
― John Belamaric, OpenStack Cloud Application Development

“Containers on a machine (or pod) must share the exact same kernel /operating system. This has repercussions if an application needs to modify the kernel, or if there is a desire to host various operating systems (or versions) on a single piece of hardware.”
― John Belamaric, OpenStack Cloud Application Development

“That being said, if performance and isolation are the upside, then efficiency and flexibility are its main downsides. Bare metal servers cannot be subdivided beyond their hardwired components. This either tends to leave a lot of underutilized hardware out there, or results in developers piggybacking multiple applications onto each physical server”
― John Belamaric, OpenStack Cloud Application Development

“Another method for scaling the database layer is to run a MySQL Cluster”
― John Belamaric, OpenStack Cloud Application Development

“One of the more popular methods for scaling the database layer is to run a Galera Cluster for MySQL”
― John Belamaric, OpenStack Cloud Application Development

“The master may not be aware of this issue and continues to operate normally. The passive instance sees the master go away and immediately puts itself into master mode. What if the other passive instance does the same exact thing? There could be three master servers all at the same time. This is commonly known as split-brain syndrome and can be a hard problem to avoid in certain failure scenarios. This problem can be even more pronounced between regions when network communication is disrupted”
― John Belamaric, OpenStack Cloud Application Development

“OpenStack also provides a Load-Balancing-as-a-Service (LBaaS), which is implemented using Neutron”
― John Belamaric, OpenStack Cloud Application Development

“The operating system installed on the instance can make a difference on how it responds to network issues as well. For example, when an instance loses a DHCP address, Ubuntu typically continues to retry renewing the DHCP address. When the network issues are resolved, the renewal process succeeds and the IP address is restored. However, RedHat and CentOS are commonly configured by default to give up after the renewal process fails, which means that even if the network issues are resolved, the instance is no longer attempting to renew the DHCP lease and permanently stays off the network”
― John Belamaric, OpenStack Cloud Application Development

“You can access the source code from our demo application via GitHub: https: / /github.com /johnbelamaric /openstack-appdev-book.”
― John Belamaric, OpenStack Cloud Application Development

“But if your organization is small and you don’t have enough resources to invest into building your own environment, you can use any CI-as-a-Service. There are two well-known services: Travis CI (https: / /travis-ci.org /) and Circle CI (https: / /circleci.com /). Feel free to pick the one you like”
― John Belamaric, OpenStack Cloud Application Development

“Applications shut down gracefully when they receive a SIGTERM signal from their manager. Unfortunately most application developers are putting worries about a graceful shutdown to the backlog”
― John Belamaric, OpenStack Cloud Application Development

“For example, if you are reusing existing ones, it is recommended that you use Python, because the OpenStack community does development and delivery for client bindings for you. If you don’t use Python, you will have to research if there are supported up-to-date client bindings or you must implement your own.”
― John Belamaric, OpenStack Cloud Application Development

“A large, single-tiered, monolithic legacy application isn’t a good fit for clouds. Efficiencies are gained when the application is modular or the load can be spread out over several application instances to allow high availability (HA) and scalability”
― John Belamaric, OpenStack Cloud Application Development

“. In short, you configure Heat and Ceilometer to monitor the Ceilometer metrics for a group of resources (say, VMs and you are monitoring CPU utilization). When a threshold is reached, an alarm fires, which in turn calls out to Heat to scale up (or down) the number of instances”
― John Belamaric, OpenStack Cloud Application Development

“Installing third-party services and applications can be difficult in any environment, but the dynamic nature of an OpenStack environment can make this problem worse. Murano is designed to solve this problem by providing an additional integration layer between third party components and the OpenStack infrastructure”
― John Belamaric, OpenStack Cloud Application Development

“For example, the ability to shift gears from Kubernetes to Swarm without having to modify your deployment system could be a huge win, and while virtual machines are likely to be a big part of the landscape for many years to come, containers are here to stay”
― John Belamaric, OpenStack Cloud Application Development