When comparing Docker Swarm vs Marathon on Apache Mesos, the Slant community recommends Docker Swarm for most people. In the question“What are the best Docker orchestration tools?” Docker Swarm is ranked 2nd while Marathon on Apache Mesos is ranked 6th. The most important reason people chose Docker Swarm is:
This also means that containers can be launched with a simple `docker run` command and Swarm will take care of the rest, such as selecting the appropriate host on which to run the container.
Ranked in these QuestionsQuestion Ranking
Pro Easy setup
This also means that containers can be launched with a simple
docker run command and Swarm will take care of the rest, such as selecting the appropriate host on which to run the container.
Pro Communicates with other Docker tools easily
Since Docker Swarm is a native Docker tool, it exposes the Docker API, making it possible t integrate it and communicate with other Docker tools (CLI, Compose, Krane, etc.).
Docker Swarm is much more lightweight than alternatives: Kubernetes and Mesosphere. Kubernetes, for instance, is very complex - it downloads and installs half of the web, where Docker Swarm has much, much smaller footprint.
Pro Compatible with Docker Compose
This gives you a well-rounded workflow.
Pro Open source
Docker Swarm is open source and provides great guides/documentation for those who want to contribute.
Pro Extremely fault-tolerant
Mesos handles (HA) master or agent failover; Marathon HA can survive scheduler failover; and Marathon automatically restarts failed tasks to maintain the desired number of task instances.
Pro Mesos is battle-tested at scale in production
Running all of Twitter, Apple's Siri, etc.
Pro Scalable to 10,000s of containers
Pro Also supports non-Docker cgroups containers
Con If a node dies in the Swarm Cluster, the containers on that node WILL be started on a another node ...
Con Bounded by the limitations of the Docker API
If the Docker API doesn't support something, then you are pretty much out of luck when it comes to Docker Swarm, because it won't be supported by Swarm either.
Con Not fault tolerant
If a node dies in your cluster, the containers on that node won't be restarted