- Spark version 1.0.9-2.1.0-1
- Configuring DC/OS Access for Spark
Run your continuous integration, automated testing, and continuous delivery jobs at scale with Jenkins for DC/OS. Instead of the static partitions so typical of other Jenkins clusters, Jenkins for DC/OS can create and destroy agents as demand increases and decreases. With multiple Jenkins masters sharing a single pool of compute resources, you can achieve much more efficient and resilient automations.
Jenkins for DC/OS runs as a root Marathon application inside a Docker container. The default Docker image contains several Jenkins plugins to get you up and running quickly. You can customize the Docker image to match your specific use case.
The Docker container also contains an NGINX reverse proxy that rewrites the URIs into the absolute paths Jenkins requires.
Jenkins for DC/OS needs a directory on disk to store its configuration and build data. To get up and running quickly, you can pin it to a single agent and use this to store the data. However, should the agent go down, your data will be lost.
Before going into production, you should set up a single POSIX-compliant file system that each agent mounts using the same path. Some options include NFS, Ceph, HDFS, and iSCSI.