![]() This takes into account the parallelism of the jobs, so a maximum of 1 job with 30x parallelism will have access to DLC per project, or 2 jobs with 15x parallelism, and so on.ĭocker_layer_caching: true # default - falseĮvery layer built in a previous job will be accessible in the Remote Docker Environment. The DLC volumes are deleted after 3 days of not being used in a job.ĬircleCI will create a maximum of 30 DLC volumes per project, so a maximum of 30 concurrent machine or Remote Docker jobs per project can have access to DLC. The volumes that are used less frequently might have older layers saved on them. This applies to parallel jobs as well: if two machine jobs are run in parallel, they will get different DLC volumes.ĭepending on which jobs the volumes are used in, they might end up with different layers saved on them. From that point on the project will have two DLC volumes associated with it. If one DLC volume exists but two jobs that request DLC are launched, CircleCI will create a new DLC volume and attach it to the second job. One DLC volume can only be attached to one machine or Remote Docker job at a time. This means that the layers downloaded in a previous job with DLC will be available in the next job that uses the same DLC volume. When the job finishes, the volume is disconnected and re-used in a future job. ![]() The volume is attached in a way that makes Docker save the image layers on the attached volume. That is, containers that are used to run your jobs are specified with the image key when using the docker executor and appear in the Spin up Environment step on your jobs pages.ĭLC is only useful when creating your own Docker image with docker build, docker compose, or similar docker commands, it does not decrease the wall clock time that all builds take to spin up the initial environment.ĭLC caches your Docker image layers by creating an external volume and attaching it to the instances that execute the machine and Remote Docker jobs. Note: DLC has no effect on Docker images used as build containers. If you are experiencing issues with cache-misses or need high-parallelism, consider trying the experimental docker-registry-image-cache orb. any job with parallelism beyond 30 will experience cache misses regardless of number of workflows running.A single job with 30 parallelism will work if only a single workflow is running, however, having more than one workflow will result in cache misses.Please note that high usage of parallelism (that is, a parallelism of 30 or above) in your configuration may cause issues with DLC, notably pulling a stale cache or no cache.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |