Storage

Let’s save your pipeline files somewhere cloud native!

When we use a Static Jenkins Server with Jenkins X we inherit the usual Jenkins storage model; that build logs and test results and reports are stored on the Persistent Volume of the Jenkins server.

However as we move towards a more Cloud Native Jenkins and use Serverless Jenkins we need a better solution for the storage of things like logs, test results, code coverage reports etc.

Storage Extensions

So we’ve added a storage extension point which is used from:

Configuring Storage

You can setup the default location to use for storage. We currently support:

  • storing files (logs, test or coverage reports) in a branch of a git repository. e.g. they could be part of your gh-pages branch for your static site.
  • storing files in Cloud Storage buckets like S3, GCS, Azure blobs etc

Storage uses classifications which are used to define the folder in which the kind of resources live such as

  • logs
  • tests
  • coverage

You can also use the special classification default which is used if you don’t have a configuration for the classification in question. e.g. you could define a location of default and then just configure where logs go if thats different.

If you are using jx boot to install and configure your setup then modify your jx-requirements.yml file to configure storage as described in the boot storage documentation

Otherwise to configure the storage location for a classification and team you use the jx edit storage

e.g.

# Configure the tests to be stored in cloud storage (using S3 / GCS / Azure Blobs etc)
jx edit storage -c tests --bucket-url s3://myExistingBucketName

# Configure the git URL and branch of where to store logs
jx edit storage -c logs --git-url https://github.com/myorg/mylogs.git --git-branch cheese

# Configure your own category
jx edit storage -c <new-category> --bucket-url gs://myExistingBucketName

You can view your teams storage settings via jx get storage

Using Stash

Inside a pipeline you can then run the jx step stash command to stash files:

# lets collect some files with the file names relative to the 'target/test-reports' folder and store in a Git URL
jx step stash -c tests -p "target/test-reports/*" --basedir target/test-reports

# lets collect some files to a specific AWS cloud storage bucket
jx step stash -c coverage -p "build/coverage/*" --bucket-url s3://my-aws-bucket
  • specify the classifier via -c such as for tests or coverage etc.
  • specify the files to collect via -p which supports wildcards like *. files which will be stored with the relative directory path
  • if you want to remove a directory prefix from the stashed files, like target/reports you can use --basedir to specify the directory to create relative file names from
NOTE Be aware that you have to run jx step stash inside your git repository, therefore dir: should be set to /workspace/source in your stash step.

By default jx step stash will use your team’s configured location for the classification you give. If you wish you can override the location for a stash using --git-url or --bucket-url

Unstashing

If you need to you can unstash previously stored files via jx step unstash

If you are in some Go source code and you have a URL from Jenkins X, such as a Build Log URL or Attachment from a PipelineActivity Custom Resource the URL could be in various forms such as:

  • gs://anotherBucket/mydir/something.txt : using a GCS bucket on GCP
  • s3://nameOfBucket/mydir/something.txt : using S3 bucket on AWS
  • azblob://thatBucket/mydir/something.txt : using an Azure bucket
  • http://foo/bar : file stored in git not using HTTPS
  • https://foo/bar : file stored in a git repo using HTTPS

If you want to easily be able to read from the URL from Go source code you can use the ReadURL function.

GKE Storage Permissions

In GKE your node-pool requires additional permissions to be able to write into GCS buckets, more specifically the devstorage.full_control permission.

If you already have a cluster, you can see these permissions in your cluster overview, under the dropdown Permissions.

The description of the field Storage should be Full, by default it is Read Only.

There are two ways to change your permissions, either you create a new cluster with the appropriate permissions or you just migrate the node-pool, if you already have a running cluster.

Create a new GKE Cluster with Full-Control

To create a new cluster with the right permissions, you need to use the --scopes flag with the storage-full argument, as seen here:

gcloud container clusters create <name> --machine-type <type> --zone <zone> --scopes=storage-full

Migrate to a node-pool with Full-Control

Migrating to a new node-pool is quite simple, it’s done in 2 steps!

First create a new node-pool with the required permissions:

gcloud container node-pools create <node-pool-name> --cluster <cluster-name> --machine-type <type> --scopes=storage-full

Now delete the old node-pool, all your Pods will be rescheduled on the new node-pool through Kubernetes magic!

gcloud container node-pools delete <node-pool-name>

Last modified April 7, 2020: release 0.0.1630 (615e14e)