{"id":70954,"date":"2020-02-23T00:58:17","date_gmt":"2020-02-22T21:58:17","guid":{"rendered":"https:\/\/prohoster.info\/blog\/ustrojstvo-helm-i-ego-podvodnye-kamni"},"modified":"2020-03-03T16:14:40","modified_gmt":"2020-03-03T13:14:40","slug":"ustrojstvo-helm-i-ego-podvodnye-kamni","status":"publish","type":"post","link":"https:\/\/prohoster.info\/en\/blog\/administrirovanie\/ustrojstvo-helm-i-ego-podvodnye-kamni","title":{"rendered":"Understanding Helm and Its Pitfalls","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"

\"Understanding
\nTyphon freight hauler concept, Anton Swanepoel<\/a><\/noindex><\/p>\n

My name is Dmitry Sugrobov, I am a developer at 'Leroy Merlin'. In this article, I will explain why Helm is needed, how it simplifies working with Kubernetes, what has changed in the third version, and how to update applications in production without downtime.<\/p>\n

This is a summary based on a presentation at the conference @Kubernetes Conference<\/a><\/noindex> by Mail.ru Cloud Solutions<\/a><\/noindex> \u2014 if you don't want to read, watch the video.
\n<\/a><\/noindex>
\n

\r\n \"Play\r\n
<\/div>\r\n <\/div><\/center><\/p>\n

Why we use Kubernetes in production<\/h2>\n

\n'Leroy Merlin' is a leader in the DIY retail market in Russia and Europe. Our company has over a hundred developers, 33,000 internal employees, and a huge number of people visiting hypermarkets and the website. To make them all happy, we decided to adhere to standard approaches in the industry. We develop new applications using microservice architecture, employ containers for environment isolation and proper delivery, and use Kubernetes for orchestration. The cost of using orchestrators is rapidly decreasing: the market is seeing more engineers skilled in the technology, and providers offering Kubernetes as a service are emerging.<\/p>\n

Everything Kubernetes does can certainly be achieved through other means, for example, by slapping scripts onto some Jenkins and docker-compose, but why complicate life when there's a ready and reliable solution? That's why we turned to Kubernetes and have been using it in production for a year now. Currently, we have twenty-four Kubernetes clusters, the oldest of which is over a year old and houses about two hundred pods.<\/p>\n

The curse of numerous YAML files in Kubernetes<\/h2>\n

\nTo deploy a microservice in Kubernetes, we create at least five YAML files: for Deployment, Service, Ingress, ConfigMap, and Secrets \u2014 and send them to the cluster. For the next application, we'll write the same set of YAML files, another one for the third application, and so on. Multiplying the number of documents by the number of environments already gives us hundreds of files, not to mention dynamic environments.<\/p>\n

\"Understanding
\nAdam Reese, core maintainer of Helm, introduced the concept of 'Development Cycle in Kubernetes<\/a><\/noindex>', which looks like this:<\/p>\n

    \n
  1. Copy YAML \u2014 copy the YAML file.<\/li>\n
  2. Paste YAML \u2014 paste it.<\/li>\n
  3. Fix Indents \u2014 fix the indents.<\/li>\n
  4. Repeat \u2014 repeat again.<\/li>\n<\/ol>\n

    \nThis method works but requires repeatedly copying YAML files. To change this cycle, Helm was invented.<\/p>\n

    What is Helm<\/h3>\n

    \nFirst of all, Helm \u2014 package manager<\/strong>, helping find and install the necessary programs. To install something like MongoDB, there's no need to visit the official website and download binaries; just execute the command helm install stable\/mongodb<\/code>.<\/p>\n

    Secondly, Helm \u2014 a templater<\/strong>, helps parameterize files. Let's return to the situation with YAML files in Kubernetes. It\u2019s easier to write the same YAML file, add some placeholders, into which Helm will substitute values. This means instead of a large set of YAML files, there will be a set of templates where the required values will be inserted at the right moment.<\/p>\n

    Thirdly, Helm \u2014 masters deployment<\/strong>. With it, applications can be installed, rolled back, and updated. Let's figure out how to do this.<\/p>\n

    \"Understanding<\/p>\n

    How to use Helm for deploying your own applications<\/h3>\n

    \nWe will install the Helm client on the computer, following the official the instructions<\/a><\/noindex>. Then we will create a set of YAML files. Instead of specifying concrete values, we\u2019ll leave placeholders that Helm will fill with information later. This set of files is called a Helm chart. There are three ways to send it to the Helm console client:<\/p>\n