Steady Integration/Steady Supply, or Steady Deployment, pipelines have been used within the software program improvement trade for years. For many groups, the times of manually taking supply code and manifest information and compiling them to create binaries or executable information after which manually distributing and putting in these purposes are lengthy gone. In an effort to automate the construct course of and distribution of software program in addition to carry out automated testing, the trade has repeatedly advanced in the direction of extra complete pipelines. Relying on how a lot of the software program improvement course of is automated, pipelines could be categorized into completely different teams and phases:
- Steady Integration is the observe of integrating code that’s being produced by builders. On medium to giant software program initiatives is frequent to have a number of builders and even a number of groups of builders work on completely different options or parts on the identical time. Taking all this code and bringing it to a central location or repository is commonly accomplished utilizing a git primarily based model management system. When the code is merged right into a department on an hourly, each day, weekly or regardless of the cadence of the event staff is, easy to advanced exams could be setup to validate the modifications and flush out potential bugs at a really early stage. When carried out in an automatic trend, all these steps consist in a steady integration pipeline.
- Steady Supply takes the pipeline to the following stage by including software program constructing and launch creation and supply. After the software program has been built-in and examined within the steady integration a part of the pipeline, steady supply provides extra testing and has the choice to deploy the newly constructed software program packages in a sandbox or stage setting for shut monitoring and extra person testing. Much like steady integration, all steps carried out within the steady supply a part of the pipeline are automated.
- Steady Deployment takes the pipeline to its subsequent and final stage. By this stage, the appliance has been built-in, examined, constructed, examined some extra, deployed in a stage setting and examined much more. The continual deployment stage takes care of deploying the appliance within the manufacturing setting. A number of completely different deployment methods can be found with completely different threat components, price issues and complexity. For instance, within the fundamental deployment mannequin, all software nodes are up to date on the identical time to the brand new model. Whereas this deployment mannequin is easy it’s also the riskiest, it’s not outage-proof and doesn’t present straightforward rollbacks. The rolling deployment mannequin because the title suggests takes an incremental strategy to updating the appliance nodes. A sure variety of nodes are up to date in batches. This mannequin supplies simpler rollback, it’s much less dangerous than the fundamental deployment however on the identical time requires that the appliance runs with each new and previous code on the identical time. In purposes that use the micro-services structure, this final requirement have to be given further consideration. A number of different deployment fashions can be found, together with canary, blue/inexperienced, A/B, and so forth.
The CI/CD pipeline part of GitLab CE
Why use CI/CD pipelines for infrastructure administration
Primarily based on the necessities of the event staff, software program improvement pipelines can take completely different types and use completely different parts. Model management techniques are often git primarily based as of late (github, gitlab, bitbucket, and so forth.). Construct and automation servers corresponding to Jenkins, drone.io, Travis CI, to call only a few, are additionally standard parts of the pipeline. The number of choices and parts make the pipelines very customizable and scalable
CI/CD pipelines have been developed and used for years and I believe it’s lastly time to think about them for infrastructure configuration and administration. The identical benefits that made CI/CD pipelines indispensable from any software program improvement enterprise apply additionally to infrastructure administration. These benefits embody:
- automation on the forefront of all steps of the pipeline
- model management and historic perception into all of the modifications
- in depth testing of all configuration modifications
- validation of modifications in a sandbox or take a look at setting previous to deployment to manufacturing
- straightforward roll-back to a identified good state in case a difficulty or bug is launched
- risk of integration with change and ticketing techniques for true infrastructure Steady Deployment
On this sequence of blogs, I’ll show how one can use Gitlab CE as a foundational part for a CI/CD pipeline that manages and configures a easy CML simulated community. A number of different parts are concerned as a part of the pipeline:
- pyATS for creating and taking snapshots of the state of the community each prior and after the modifications have been utilized
- Ansible for performing the configuration modifications
- Cisco CML to simulate a 4 node community that can act because the take a look at infrastructure
Easy community simulation in Cisco CML
Keep tuned for a deeper dive
Subsequent up on this weblog sequence we’ll dive deeper into Gitlab CE, and the CI/CD pipeline part. If in case you have any questions or feedback, please depart me a remark within the part under.
Within the meantime, if you want to study and observe among the newest methods, please go to our Cisco Modeling Labs and the CML Sandbox. A world of community simulation awaits!
We’d love to listen to what you suppose. Ask a query or depart a remark under.
And keep related with Cisco DevNet on social!