.circleci

Setting CI/CD system in CircleCI

Here are the steps requires to start building this project at your own space:

1. Fork the project to your organization;
2. Setup account CircleCI.com;
3. Link GitHub to CircleCI and select forked project;
4. Press "Start building;
5. Skip configuration file insertion as it is already inserted;
6. Go to https://app.circleci.com/settings/project/github/<YOUR_ORGANIZATION>/filecoin-cloud-images/environment-variables and setup environment variables according to the environment variables section;
7. Go to the main project page and approve building of those pipelines that are marked as <PROVIDER>-base;
8. You can schedule automatic incremental image recreation. To do that - go to .circleci/config.yml and uncomment sections like:

          #    triggers:
          #      - schedule:
          #          cron: "0 0 * * *"
          #          filters:
          #            branches:
          #              only:
          #                - master

9. Next builds will be triggered at 12.00 AM on daily basis. You can adjust it by changing cron: "0 0 * * *" this line in accordance with Linux Cron system syntax;
10. A number of public settings such as disk size and AMI regions you can control by direct editing of Packer configuration files inside of each provider folder.

Environment variables

AWS provider-related variables

• AWS_ACCESS_KEY;
• AWS_SECRET_KEY.

To generate access and secret keys visit "My security credentials" page and press "Create access key". Save the keypair and enter as separate environment variables.

Azure provider-related variables

• AZURE_CLIENT_ID;
• AZURE_CLIENT_SECRET;
• AZURE_TENANT_ID.

To generate client's authenthication data do the following:

1. Go to the App registrations panel of the Azure Active Directory and create a new application. This is an Azure version of Service Accounts;
2. Open the newly registered app. At the Overview tab save the Application (client) ID and Directory (tenant) ID values;
3. Go to the Certificates and Secretes tab. Press New client secret to generate the secret.

• AZURE_SUBSCRIPTION_ID - open the Subscriptions blade and copy the ID of the subscription which you would like to use in this project;
• AZURE_RESOURCE_GROUP_NAME - you will need a separate resource group to create resources at. Go to resource group management dashboard and click Add to create a new resource group;
• AZURE_STORAGE_ACCOUNT;
• AZURE_STORAGE_ACCOUNT_CONTAINER.

This solution will create a .vhd disks as an output and store them at Azure storage account. Thus, you will need to provide scripts with storage account identification data:

1. Go to the Azure storage accounts management dashboard;
2. Click "Add" to create new storage account. Make it public or private according to your needs;
3. Open created storage account and select Containers;
4. Press + Container to add a new container for VHDs;
5. Set the name of the azure storage account as AZURE_STORAGE_ACCOUNT and container name as AZURE_STORAGE_ACCOUNT_CONTAINER.

GCP provider-related variables

• GCP_ACCOUNT - JSON one-liner that contains authenthication information for GCP.

To get JSON do the following:

1. Open the Service Accounts tab at GCP IAM;
2. Press "Create service account";
3. Set the name and click Create;
4. Press ⋮ button which is on the same line as the created service account;
5. Press Create key. Select JSON and hit Create;
6. Go to IAM panel and assign the Compute Instance Admin (v1), Service Account User and Storage Admin roles to the created service account;
7. Convert saved JSON file into one-liner. For that purpose you can use jq. For example: jq -c . account.json.

• GCP_PROJECT - the project where images will be built. The account specified at GCP_ACCOUNT variable must have access to the Compute API. Compute API must be enabled;
• GCP_STORAGE - storage account where images will be published. The account specified at GCP_ACCOUNT variable must have access to the Storage API. Storage API must be enabled.

Filecoin-related variables

• BRANCHES (comma-separated values) - set of branches that should be searched for tags to build Lotus node from. Currently we are building interopnet and master branches, which stands for interop and test networks. The variable's value in our case will be equal to interopnet,master.

Optional variables:

• SLACK_WEBHOOK - URL to the webhook that allows posting building information to the slack channels. Note: To make build end successfully - comment out - slack/status lines if you haven't set the SLACK_WEBHOOK variables.