docs(mesos): describe Marathon installation and integration with deis

smothiki · smothiki · commit 8dc82e6bc461 · 2015-07-20T16:16:21.000-06:00
diff --git a/controller/requirements.txt b/controller/requirements.txt
@@ -18,4 +18,4 @@ setproctitle==1.1.8
 static==1.1.1
 South==1.0.2
 python-ldap==2.4.19
-marathon
+marathon==0.6.15
diff --git a/deisctl/cmd/cmd.go b/deisctl/cmd/cmd.go
@@ -75,8 +75,7 @@ func Start(targets []string, b backend.Backend) error {
 			return StartPlatform(b, true)
 		} else if targets[0] == swarm {
 			return StartSwarm(b)
-		}
-		if targets[0] == mesos {
+		} else if targets[0] == mesos {
 			return StartMesos(b)
 		}
 	}
@@ -184,8 +183,7 @@ func Stop(targets []string, b backend.Backend) error {
 			return StopPlatform(b, true)
 		} else if targets[0] == swarm {
 			return StopSwarm(b)
-		}
-		if targets[0] == mesos {
+		} else if targets[0] == mesos {
 			return StopMesos(b)
 		}
 	}
@@ -289,8 +287,7 @@ func Install(targets []string, b backend.Backend, checkKeys func() error) error
 			return InstallPlatform(b, checkKeys, true)
 		} else if targets[0] == swarm {
 			return InstallSwarm(b)
-		}
-		if targets[0] == mesos {
+		} else if targets[0] == mesos {
 			return InstallMesos(b)
 		}
 	}
@@ -353,8 +350,7 @@ func Uninstall(targets []string, b backend.Backend) error {
 			return UninstallPlatform(b, true)
 		} else if targets[0] == swarm {
 			return UnInstallSwarm(b)
-		}
-		if targets[0] == mesos {
+		} else if targets[0] == mesos {
 			return UninstallMesos(b)
 		}
 	}
diff --git a/deisctl/units/deis-mesos-marathon.service b/deisctl/units/deis-mesos-marathon.service
@@ -10,15 +10,15 @@ RestartSec=20
 TimeoutStartSec=0
 ExecStartPre=-/bin/sh -c "etcdctl get /deis/scheduler/mesos/marathon >/dev/null 2>&1 || etcdctl mk /deis/scheduler/mesos/marathon"
 ExecStartPre=/bin/sh -c "etcdctl set /deis/scheduler/mesos/marathon $COREOS_PRIVATE_IPV4"
-ExecStartPre=-/usr/bin/docker kill coreos-mesos-marathon
-ExecStartPre=-/usr/bin/docker rm coreos-mesos-marathon
-ExecStartPre=/usr/bin/docker pull aledbf/mesos-marathon:0.22.1
-ExecStart=/usr/bin/sh -c "/usr/bin/docker run \
+ExecStartPre=/bin/sh -c "IMAGE=`/run/deis/bin/get_image /deis/mesos-marathon` && docker history $IMAGE >/dev/null 2>&1 || docker pull $IMAGE"
+ExecStartPre=-/usr/bin/docker kill deis-mesos-marathon
+ExecStartPre=-/usr/bin/docker rm deis-mesos-marathon
+ExecStart=/usr/bin/sh -c "IMAGE=`/run/deis/bin/get_image /deis/mesos-marathon` && docker run \
   --name=coreos-mesos-marathon \
   --net=host \
   -e HOST=$COREOS_PRIVATE_IPV4 \
-  aledbf/mesos-marathon:0.22.1"
-ExecStop=-/usr/bin/docker stop coreos-mesos-marathon
+  $IMAGE"
+ExecStop=-/usr/bin/docker stop deis-mesos-marathon
 
 [Install]
 WantedBy=multi-user.target
diff --git a/deisctl/units/deis-mesos-master.service b/deisctl/units/deis-mesos-master.service
@@ -8,18 +8,18 @@ EnvironmentFile=/etc/environment
 Restart=on-failure
 RestartSec=20
 TimeoutStartSec=0
-ExecStartPre=-/usr/bin/docker kill coreos-mesos-master
-ExecStartPre=-/usr/bin/docker rm coreos-mesos-master
-ExecStartPre=/bin/sh -c "docker inspect coreos-mesos-master-data >/dev/null 2>&1 || docker run --name coreos-mesos-master-data -v /tmp/mesos-master alpine:3.1 /bin/true"
-ExecStartPre=/usr/bin/docker pull aledbf/mesos-master:0.22.1
-ExecStart=/usr/bin/sh -c "/usr/bin/docker run \
-  --volumes-from=coreos-mesos-master-data \
-  --name=coreos-mesos-master \
+ExecStartPre=-/usr/bin/docker kill deis-mesos-master
+ExecStartPre=-/usr/bin/docker rm deis-mesos-master
+ExecStartPre=/bin/sh -c "docker inspect deis-mesos-master-data >/dev/null 2>&1 || docker run --name deis-mesos-master-data -v /tmp/mesos-master alpine:3.1 /bin/true"
+ExecStartPre=/bin/sh -c "IMAGE=`/run/deis/bin/get_image /deis/mesos-master` && docker history $IMAGE >/dev/null 2>&1 || docker pull $IMAGE"
+ExecStart=/usr/bin/sh -c "IMAGE=`/run/deis/bin/get_image /deis/mesos-master` && docker run \
+  --volumes-from=deis-mesos-master-data \
+  --name=deis-mesos-master \
   --privileged \
   --net=host \
   -e HOST=$COREOS_PRIVATE_IPV4 \
-  aledbf/mesos-master:0.22.1"
-ExecStop=-/usr/bin/docker stop coreos-mesos-master
+  $IMAGE"
+ExecStop=-/usr/bin/docker stop deis-mesos-master
 
 [Install]
 WantedBy=multi-user.target
diff --git a/deisctl/units/deis-mesos-slave.service b/deisctl/units/deis-mesos-slave.service
@@ -8,19 +8,19 @@ EnvironmentFile=/etc/environment
 Restart=on-failure
 RestartSec=20
 TimeoutStartSec=0
-ExecStartPre=-/usr/bin/docker kill coreos-mesos-slave
-ExecStartPre=-/usr/bin/docker rm coreos-mesos-slave
-ExecStartPre=/usr/bin/docker pull aledbf/mesos-slave:0.22.1
-ExecStart=/usr/bin/sh -c "/usr/bin/docker run \
-  --name=coreos-mesos-slave \
+ExecStartPre=-/usr/bin/docker kill deis-mesos-slave
+ExecStartPre=-/usr/bin/docker rm deis-mesos-slave
+ExecStartPre=/bin/sh -c "IMAGE=`/run/deis/bin/get_image /deis/mesos-slave` && docker history $IMAGE >/dev/null 2>&1 || docker pull $IMAGE"
+ExecStart=/usr/bin/sh -c "IMAGE=`/run/deis/bin/get_image /deis/mesos-slave` && docker run \
+  --name=deis-mesos-slave \
   --net=host \
   --privileged \
   -e HOST=$COREOS_PRIVATE_IPV4 \
   -v /sys:/sys \
   -v /usr/bin/docker:/usr/bin/docker:ro \
   -v /var/run/docker.sock:/var/run/docker.sock \
   -v /lib64/libdevmapper.so.1.02:/lib/libdevmapper.so.1.02:ro \
-  aledbf/mesos-slave:0.22.1"
+  $IMAGE"
 ExecStop=/usr/bin/docker stop coreos-mesos-slave
 
 [Install]
diff --git a/deisctl/units/deis-mesos-zk.service b/deisctl/units/deis-mesos-zk.service
@@ -9,11 +9,11 @@ Restart=on-failure
 RestartSec=20
 TimeoutStartSec=0
 ExecStartPre=/bin/sh -c "docker inspect zookeeper-data >/dev/null 2>&1 || docker run --name zookeeper-data -v /opt/zookeeper-data alpine:3.1 /bin/true"
-ExecStartPre=-/usr/bin/docker kill zookeeper
-ExecStartPre=-/usr/bin/docker rm zookeeper
-ExecStartPre=/usr/bin/docker pull aledbf/zookeeper:3.5.0
-ExecStart=/bin/sh -c "/usr/bin/docker run -e EXTERNAL_PORT=2181 -e HOST=$COREOS_PRIVATE_IPV4 -e LOG_LEVEL=debug --net=host --rm --name zookeeper --volumes-from=zookeeper-data aledbf/zookeeper:3.5.0"
-ExecStop=/usr/bin/docker stop zookeeper
+ExecStartPre=-/usr/bin/docker kill deis-zookeeper
+ExecStartPre=-/usr/bin/docker rm deis-zookeeper
+ExecStartPre=/bin/sh -c "IMAGE=`/run/deis/bin/get_image /deis/zookeeper` && docker history $IMAGE >/dev/null 2>&1 || docker pull $IMAGE"
+ExecStart=/bin/sh -c "IMAGE=`/run/deis/bin/get_image /deis/zookeeper` && docker run -e EXTERNAL_PORT=2181 -e HOST=$COREOS_PRIVATE_IPV4 -e LOG_LEVEL=debug --net=host --rm --name deis-zookeeper --volumes-from=zookeeper-data $IMAGE"
+ExecStop=/usr/bin/docker stop deis-zookeeper
 
 [Install]
 WantedBy=multi-user.target
diff --git a/docs/customizing_deis/choosing-a-scheduler.rst b/docs/customizing_deis/choosing-a-scheduler.rst
@@ -12,9 +12,14 @@ of your app. For example, a command such as ``deis scale web=3`` tells the
 scheduler to run three containers from the Docker image for your app.
 
 Deis defaults to using the `Fleet Scheduler`_. A tech preview of the `Swarm Scheduler`_
-is available for testing. Work is ongoing on `Kubernetes`_ and `Mesos`_-based
-schedulers with the intent to test those alternatives in future releases of Deis.
+and `Mesos with Marathon framework`_ is available for testing. Work is ongoing on `Kubernetes`_-based
+scheduler with the intent to test those alternatives in future releases of Deis. Keep watching `dynamic metadata fleet PR 1077`_
+for support of runtime changes to metadata.
 
+.. note::
+
+    If you are using a scheduler other than fleet app containers will not be rescheduled if deis-registry is unavailable.
+    For more information look at `deis-registry issue 3619`.
 
 Settings set by scheduler
 -------------------------
@@ -23,12 +28,13 @@ The following etcd keys are set by the scheduler module of the controller compon
 
 Some keys will exist only if a particular ``schedulerModule`` backend is enabled.
 
-=============================            ================================================
-setting                                  description
-=============================            ================================================
-/deis/scheduler/swarm/host               the swarm manager's host IP address
-/deis/scheduler/swarm/node               used to identify other nodes in the cluster
-=============================            ================================================
+===================================            ======================================================
+setting                                        description
+===================================            ======================================================
+/deis/scheduler/swarm/host                     the swarm manager's host IP address
+/deis/scheduler/swarm/node                     used to identify other nodes in the cluster
+/deis/scheduler/mesos/marathon                 used to identify Marathon framework's host IP address
+===================================            ======================================================
 
 
 Settings used by scheduler
@@ -39,8 +45,8 @@ The following etcd keys are used by the scheduler module of the controller compo
 ====================================      ===============================================
 setting                                   description
 ====================================      ===============================================
-/deis/controller/schedulerModule          scheduler backend, either "fleet" or "swarm"
-                                          (default: "fleet")
+/deis/controller/schedulerModule          scheduler backend, either "fleet" or "swarm" or
+                                          mesos_marathon (default: "fleet")
 ====================================      ===============================================
 
 
@@ -97,10 +103,7 @@ to `enable the remote API`_.
     **Known Issues**
 
     - It is not yet possible to change the default affinity filter.
-    - If swarm can't create all the containers requested, Deis returns an error
-      and leaves some containers in the "created" state. Until this behavior is fixed,
-      be mindful of resource limitations on your cluster.
-    - App containers will not be rescheduled if deis-registry is unavailable.
+    - If swarm can't create all the containers requested during scale, deis rolls back the scale operation.
 
 To test the Swarm Scheduler backend, first install and start the swarm components:
 
@@ -132,9 +135,68 @@ component, or spy on Docker events directly on the swarm-manager machine:
     2015-04-30T17:31 61e59c: (from 172.17.8.100:5000/hungry-variable:v5 node:deis-02) start
 
 
+Mesos with Marathon framework
+-----------------------------
+
+.. important::
+
+    The Mesos with Marathon framework Scheduler is a technology preview and is not recommended for
+    production use.
+
+`Mesos`_ is a distributed system kernel:
+
+    Mesos provides APIs for resource management and scheduling. A framework interacts with Mesos master
+    and schedules and task. A Zookeeper cluster elects Mesos master node. Mesos slaves are installed on
+    each node and they communicate to master with available resources.
+
+`Marathon`_ is a Mesos_ framework for long running applications:
+
+    Marathon provides a Paas like feel for long running applications and features like high-availablilty, host constraints,
+    service discovery, load balancing and REST API to control your Apps.
+
+Deis uses the Marathon framework to schedule containers. Since Marathon is a framework for long-running
+jobs, Deis uses the `Fleet Scheduler`_ to run batch processing jobs. ``deisctl`` installs a standalone Mesos
+cluster. To install an HA Mesos cluster, follow the directions at `aledbf-mesos`_, and set the etcd key
+``/deis/scheduler/mesos/marathon`` to any Marathon node IP address. If a request is received by a regular
+Marathon node, it is proxied to the master Marathon node.
+
+To test the Marathon Scheduler backend, first install and start the mesos components:
+
+.. code-block:: console
+
+    $ deisctl install mesos && deisctl start mesos
+
+Then set the controller's ``schedulerModule`` to "mesos_marathon":
+
+.. code-block:: console
+
+    $ deisctl config controller set schedulerModule=mesos_marathon
+
+The Marathon framework is now active. Commands such as ``deis destroy`` or
+``deis scale web=9`` will use `Marathon`_ to manage app containers.
+
+Deis starts Marathon on port 8180. You can manage apps through the Marathon UI, which is accessible at http://<Marathon-node-IP>:8180
+
+.. note::
+
+    **Known Issues**
+
+    - deisctl installs a standalone mesos cluster as fleet doesn't support runtime change to metadata.
+      You can specify this in cloud-init during the deployment of the node. keep watching `dynamic metadata fleet PR 1077`_.
+    - If you want to access Marathon UI, you'll have to expose port 8180 in the security group settings.
+      This is blocked off by default for security purposes.
+    - Deis does not yet use Marathon's docker container API to create containers.
+    - CPU shares are integers representing the number of CPUs. Memory limits should be specified in MB.
+
+
 .. _Kubernetes: http://kubernetes.io/
-.. _Mesos: http://mesos.apache.org/
+.. _Mesos: http://mesos.apache.org
+.. _Marathon: https://github.com/mesosphere/marathon
 .. _fleet: https://github.com/coreos/fleet#fleet---a-distributed-init-system
 .. _swarm: https://github.com/docker/swarm#swarm-a-docker-native-clustering-system
 .. _`soft affinity`: https://docs.docker.com/swarm/scheduler/filter/#soft-affinitiesconstraints
 .. _`enable the remote API`: https://coreos.com/docs/launching-containers/building/customizing-docker/
+.. _`deis-kubernetes issue 3850`: https://github.com/deis/deis/issues/3850
+.. _`dynamic metadata fleet PR 1077`: https://github.com/coreos/fleet/pull/1077
+.. _`aledbf-mesos`: https://github.com/aledbf/coreos-mesos-zookeeper
+.. _`deis-registry issue 3619`: https://github.com/deis/deis/issues/3619

Original file line number	Diff line number	Diff line change
`@@ -75,8 +75,7 @@ func Start(targets []string, b backend.Backend) error {`
`75`	`75`	`return StartPlatform(b, true)`
`76`	`76`	`} else if targets[0] == swarm {`
`77`	`77`	`return StartSwarm(b)`
`78`		`- }`
`79`		`- if targets[0] == mesos {`
	`78`	`+ } else if targets[0] == mesos {`
`80`	`79`	`return StartMesos(b)`
`81`	`80`	`}`
`82`	`81`	`}`
`@@ -184,8 +183,7 @@ func Stop(targets []string, b backend.Backend) error {`
`184`	`183`	`return StopPlatform(b, true)`
`185`	`184`	`} else if targets[0] == swarm {`
`186`	`185`	`return StopSwarm(b)`
`187`		`- }`
`188`		`- if targets[0] == mesos {`
	`186`	`+ } else if targets[0] == mesos {`
`189`	`187`	`return StopMesos(b)`
`190`	`188`	`}`
`191`	`189`	`}`
`@@ -289,8 +287,7 @@ func Install(targets []string, b backend.Backend, checkKeys func() error) error`
`289`	`287`	`return InstallPlatform(b, checkKeys, true)`
`290`	`288`	`} else if targets[0] == swarm {`
`291`	`289`	`return InstallSwarm(b)`
`292`		`- }`
`293`		`- if targets[0] == mesos {`
	`290`	`+ } else if targets[0] == mesos {`
`294`	`291`	`return InstallMesos(b)`
`295`	`292`	`}`
`296`	`293`	`}`
`@@ -353,8 +350,7 @@ func Uninstall(targets []string, b backend.Backend) error {`
`353`	`350`	`return UninstallPlatform(b, true)`
`354`	`351`	`} else if targets[0] == swarm {`
`355`	`352`	`return UnInstallSwarm(b)`
`356`		`- }`
`357`		`- if targets[0] == mesos {`
	`353`	`+ } else if targets[0] == mesos {`
`358`	`354`	`return UninstallMesos(b)`
`359`	`355`	`}`
`360`	`356`	`}`