Upgrade Testing with Enos

We're using Enos to perform upgrade testing. These tests are run via GitHub Actions from the private nomad-e2e repository. This document describes how you can run these tests from your local development environment if you're a HashiCorp developer.

There are two major components to be aware of:

• This directory includes the upgrade scenario and the Terraform modules and shell scripts needed to execute that scenario.
• The scenario uses the same cluster provisioning infrastructure as the E2E tests in the e2e/ directory in the root of this repo. So to run the upgrade scenario you also have to have all the credentials set up to run the E2E tests. (We may try to fold these together in the future.)

The terraform/ folder has provisioning code to spin up a Nomad cluster on AWS. You'll need both Terraform and AWS credentials to setup AWS instances on which e2e tests will run. See the README for details. The number of servers and clients is configurable, as is the specific build of Nomad to deploy and the configuration file for each client and server.

Setup

You'll need a recent version of Terraform, the most current version of Enos, as well as AWS credentials to create the Nomad cluster and credentials for HCP. The Terraform configurations assume that an appropriate instance role has been configured elsewhere and that you have the ability to AssumeRole into the AWS account.

Configure the following environment variables. For HashiCorp Nomad developers, this configuration can be found in 1Pass in the Nomad team's vault under nomad-e2e.

export HCP_CLIENT_ID=
export HCP_CLIENT_SECRET=

The Vault admin token will expire after 6 hours. If you haven't created one already use the separate Terraform configuration found in the $REPO/e2e/terraform/hcp-vault-auth directory. The following will set the correct values for VAULT_TOKEN, VAULT_ADDR, and VAULT_NAMESPACE:

terraform init
terraform apply --auto-approve
$(terraform output --raw environment)

Make sure your AWS credentials have been refreshed with the appropriate IAM role:

$ doormat login --force
$ doormat aws cred-file add-profile --role "$ROLE" --set-default

Next you'll need to obtain an Artifactory token via Doormat.

export ARTIFACTORY_TOKEN=$(doormat artifactory create-token | jq -r .access_token)

Next you'll need to populate the Enos variables file `enos.vars.hcl (unlike Terraform, Enos doesn't accept variables on the command line):

artifactory_username = "<your email address>"
artifactory_token    = "<your ARTIFACTORY_TOKEN from above>"
product_version      = "1.8.9"                        # starting version
upgrade_version      = "1.9.4"                        # version to upgrade to
download_binary_path = "/home/foo/Downloads/nomad"    # directory on your machine to download binaries
nomad_license        = "<your Nomad Enterprise license, when running Nomad ENT>"
consul_license       = "<your Consul Enterprise license, currently always required>"
aws_region           = "us-east-1"

When the variables file is placed in the enos root folder with the name enos.vars.hcl it is automatically picked up by enos, if a different variables files will be used, it can be pass using the flag --var-file.

Reviewing Enos

You can quickly validate the Enos scenario configuration without running it:

$ enos scenario validate upgrade --var-file /tmp/enos.vars
$ echo $?
0

You can also review what Enos will do by generating an outline you can read in your browser:

$ enos scenario outline upgrade --var-file /tmp/enos.vars --format=html > /tmp/outline.html
$ open /tmp/outline.html

Running Enos

Run the Enos scenario end-to-end:

$ enos scenario run upgrade --var-file /tmp/enos.vars --timeout 2h

Enos will not clean up after itself automatically if interrupted. If you have to interrupt it, you may need to run enos scenario destroy upgrade --var-file /tmp/enos.vars

Debugging

Enos builds Terraform state in the .enos directory, in a subdirectory named with a hash. If you're working on Enos scenarios or test workloads and want to connect to the Nomad cluster you create, you can use the debug-environment script in this directory to set your Nomad environment variables by passing it the path to that subdirectory. For example:

$ $(./debug-environment .enos/c545bbc25c5eec0ca86c99595a9034b5451a91aa10b586da2baab435df65be2e)

Note that this won't be fully populated until the Enos scenario is far enough along to bootstrap the Nomad cluster.

Adding New Workloads

As part of the testing process some test workloads are dispatched and are expected to run during all the update process, they are stored under enos/modules/run_workloads/jobs and must be defined with the following attributes:

Required Attributes

• job_spec (string): Path to the job specification for your workload. The path should be relative to the run_workloads module. For example: jobs/raw-exec-service.nomad.hcl.

• alloc_count (number): This variable serves two purposes:

  1. Every workload must define the alloc_count variable, regardless of whether it is actively used. This is because jobs are executed using this command:

     variable "alloc_count" {
       type = number
     }

  This is done to force the job spec author to add a value to the alloc_count. 2. It is used to calculate the expected number of allocations in the cluster once all jobs are running.

  If the variable is missing or left undefined, the job will fail to run, which will impact the upgrade scenario.

  For system jobs, the number of allocations is determined by the number of nodes. In such cases, alloc_count is conventionally set to 0, as it is not directly used.

• type (string): Specifies the type of workload—service, batch, or system. Setting the correct type is important, as it affects the calculation of the total number of expected allocations in the cluster.

Optional Attributes

The following attributes are only required if your workload has prerequisites or final configurations before it is fully operational. For example, a job using tproxy may require a new intention to be configured in Consul.

• pre_script (optional, string): Path to a script that should be executed before the job runs.
• post_script (optional, string): Path to a script that should be executed after the job runs.

All scripts are located in enos/modules/run_workloads/scripts. Similar to job_spec, the path should be relative to the run_workloads module. Example: scripts/wait_for_nfs_volume.sh.

Adding a New Workload

If you want to add a new workload to test a specific feature, follow these steps:

1. Modify the run_initial_workloads step in enos-scenario-upgrade.hcl and include your workload in the workloads variable.

2. Add the job specification and any necessary pre/post scripts to the appropriate directories:

   • enos/modules/run_workloads/jobs
   • enos/modules/run_workloads/scripts

Important:

• Ensure that the alloc_count variable is included in the job specification. If it is missing or undefined, the job will fail to run, potentially disrupting the upgrade scenario.

• During normal execution of the test and to verify the health of the cluster, the number of jobs and allocs running is verified multiple times at different stages of the process. Make sure your job has a health check, to ensure it will be restarted in case of unexpected failures and if it is a batch job, it will not exit before the test has concluded.

If you want to verify your workload without having to run all the scenario, you can manually pass values to variables with flags or a .tfvars file and run the module from the run_workloads directory like you would any other terraform module.