Download and cache a custom file#
This guide shows you how to download one or more files and store them in Semaphore cache indefinitely.
Example Semaphore project#
.semaphore/semaphore.yml file of the project is shown below:
version: v1.0 name: Cache custom file agent: machine: type: e1-standard-2 os_image: ubuntu1804 blocks: - name: Use cache or create task: jobs: - name: Get .deb from cache or create cache commands: - checkout - source .semaphore/get_packages.sh - ls -l packages - name: Reuse from cache task: prologue: commands: - checkout - cache restore $SEMAPHORE_PROJECT_NAME-deps - sudo dpkg -i ./packages/enscript.deb
get_packages.sh script referenced above has the following content:
cache restore $SEMAPHORE_PROJECT_NAME-deps if [ -d 'packages' ]; then echo 'Found packages in cache' else mkdir packages wget http://ge.archive.ubuntu.com/ubuntu/pool/universe/e/enscript/enscript_184.108.40.206-3_amd64.deb -O ./packages/enscript.deb # (add more downloads here if needed) cache store $SEMAPHORE_PROJECT_NAME-deps packages fi
In the example above, the deb package will be downloaded from the external source only once, and reused from the cache in all subsequent workflows.
You can extend this example to download any additional files and store them
together in the
SEMAPHORE_PROJECT_NAMEis one of the environment variables that is available in all blocks of a pipeline.
Semaphore cache is built on the idea of speed over high-availability. Data availability is not 100% guaranteed and Semaphore isn't responsible for data loss related to the cache. For this reason, cached files may not always be available in your jobs.
cacheutility returns exit code 0 in the event of unsuccessful writes and reads from the cache store. If the cache is not available, your pipeline should be able to recover — the same way caching works in a web application, for example.