digitalearns / upswing Goto Github PK

Ansible 2.1 is out.
It's good if everyone could upgrade their ansible version

sudo -H pip install --upgrade ansible

How about going the Google Hangouts path for all future meetups for DigitalEarns. The two advantages being, people need not be present at the same place and there will be option to record the hangout so that people absent can see it later.

This issue can be used as a place to discuss the steps we need to take towards implementing TLS for docker swarm.

The primary reading material for this is the official documentation https://docs.docker.com/swarm/configure-tls/

Please add if any better readings are available.

Create Ansible role for docker-swarm installation.

We will consider deploying the whole stack on to a different environment other than vagrant to test. DigitalOcean is a preferred option.

We will be using Vagrant for all our local development and test purpose. Vagrantfile will require multi-machine configuration with a suitable IP range. The basebox should be the official CentOS 7 vagrant image. 4 VM's are planned in the initial stage, this may change going forward.

How about creating multiple branches to the master like infrastructure dev etc.? so that those who want to start working on the application to be deployed can do it.

This is not of high priority. But it is good to have a custom test data generator (also can explore other open source repos which can be reused) to enable easy testing for users.

We will be using Ansible as the preferred configuration management tool.

Write Ansible role for Docker Installation. Try to make the role generic. As we are using CentOS 7, make sure to create a drop-in systemd unit file custom docker daemon configurations. If possible cover the multi-host networking configurations too.

Note : We wont be shipping installers as part of the docker roles.

We will have different types of projects as candidates to undergo testing on this environment. Scalding-Scala-SBT project with unit tests(local mode on scalding) is one of them.

Need to define various milestones. To discuss with @Ranjandas and come up with some targets.

As we have decided to deploy all the services as docker containers we should have a cluster'ing solution which can take care of the containers across multiple nodes. A solution that can address scaling, load-balancing, and availability of the services. As we have the native docker clustering solution production ready as of November last year, we will stick to it for requirement. Docker swarm is simple compared to various other cluster manager solutions. If it doesn't fulfil our needs we will look for other options.

Integrate Docker Swarm to the 4 Node Docker environment with Multi-Host Networking.

It could be a while to have the set up for various OS users(I assume). But it is good to mention in the README, the type of OS and its prerequisites, s/w libraries to be installed etc for developers to start with.

Should we start using waffle.io or something similar?

Document the details involved in the project. Decide on a suitable way of documenting stuffs.

Various options are

• to create .md files inside a docs directory
• to maintain github wiki page
• to use github-pages and maintain a small website giving status updates

R is of much interests to many parties. Environment should provide capabilities to run R pipeline jobs.

Are we planning to use the same repo for all kind of code say for instance for #12 are we planning to manage it in the same repo?

I would suggest the owner of the same can create a separate repo and others following the task could contribute to it, this will give more among us an opportunity to own a repo.

Attendees should be able to do basic docker operations and should be able to understand and play around with Docker Multi-Host Networking.

Create Ansible role for NTP configuration across all the servers. Proper time sync is very important in distributed systems.

README should be simple enough for people to understand what the repo does and how to use it. The picture @Ranjandas shared in gitter was nice.

As you know for this project we are using CentOS 7 and the firewall services has been moved from iptables to FirewallD. This task is for people interested in security space to explore firewalld and learn how firewallD can be used with ansible. It would be nice to have a knowledge sharing session on firewallD too.

Even though we initially will not be enforcing strict security policies , we should definitely consider security primary.

Take knowledge sharing session on Docker Swarm.

As all of us wont be working on same stuff its very important to pass on what we have learned to others.

Take a session on Vagrant and various configurations options available. Cover configuration options that is seen by default in a standard Vagrantfile.

There are very few candidates available for Static Code Analysis and Code Coverage tools available for Scala, please suggest which we can use that integrates with the build focusing on CI/CD.

Eg:

• Codacy
• SCoverage
• ScapeGoat

Are we going to use the official docker images of different components or build everything from scratch?

We are planning to deploy all our services in the CI/CD pipeline as Docker containers. This will require a stable pool of Docker enabled machines that can seamlessly host docker containers. We will be also giving importance to Docker Multi-Host networking as this will enable containers on different host to talk to each other. As docker now has native multi-host networking capabilities, we are planning to use the same throughout this project.

Install Docker Engine on 4 nodes and enable multi-host networking. Test the connectivity of containers from all the hosts.

Looks like Red Hat and Netflix are two major users of PCP. http://pcp.io/. Lets evaluate this for upswing.