Skip to content

Latest commit

 

History

History
104 lines (76 loc) · 6.08 KB

INSTALL.md

File metadata and controls

104 lines (76 loc) · 6.08 KB

oktoflow platform: Installation

This document summarizes how to install and run the oktoflow platform. We discuss the usual installation from scratch but also the use of the pre-package example containers on Dockerhub.

Installation from scratch

We recommend a full installation from scratch using the installation package. You can obtain the actual platform install package for Windows and for Linux.

Unpack the installation package, e.g., on Linux

tar xzvf install.tar.gz

All installation steps are described in the README of the installation package, which is also contained in the respective archive files. The required software installation is detailed in PREREQUISITES.

Pre-packaged Docker containers

For a simple startup, we offer two pre-packaged Docker containers on Docker Hub. These containers are meant

  • containing the platform parts (AAS server, the ECS Runtime, the Service Manager) and a simple example service artifact.
  • one with the Command Line Interface shown in the usage scenario above.

To experiment with the containers, use the following commands. Create a network:

docker network create --subnet=172.19.0.0/16 platform

and start the platform parts:

docker run --network platform --ip 172.19.0.22 -p 9001:9001 -p 9002:9002 -p 8883:8883 -p 9090:9090 -p 9091:9091 -p 4200:4200 iipecosphere/platform:platform_all.latest

and similarly for the CLI

docker run -i --network platform iipecosphere/platform:cli.latest

The service artifact that is needed to add/start services is available under the following local URI [the name is changing from SimpleMeshTestingApp-0.1.0-SNAPSHOT.jar to SimpleMeshTestingApp-0.1.0-SNAPSHOT-bin.jar depending on your version]

file:/device/SimpleMeshTestingApp-0.1.0-SNAPSHOT.jar

This setup uses port 8883 for the AMQP broker, 9001 for the Platform Asset Administration Shell (AAS) and 9002 for the AAS registry.

We discuss now a simple scenario on how the platform can be used through the command line interface:

oktoflow, interactive platform command line
AAS server: http://127.0.0.1:9001
AAS registry: http://127.0.0.1:9002/registry
Type "help" for help.
> resources
resources> help
 list
 help
 back
resources> list
 - Resource a005056C00008
   systemdisktotal: 1023887356
   systemmemorytotal: 2147483647
   simplemeterlist: ["system.cpu.count","system.cpu.usage", "system.disk.free", "system.memory.free"…]
   containerSystemName: Docker
   systemmemoryfree: 2147483647
   systemdiskfree: 464061712
   systemmemoryused: 2147483647
   systemdiskusable: 464061712
   systemmemoryusage: 0.5555296172875698
   systemdiskused: 559825644
resources> back
> exit

The trace above illustrates a shell-like interaction of the user with the command line interface. The CLI provides commands for resources, containers and services. Resources are added automatically when an ECS runtime is executed. The trace above shows the properties listing for a resource with identifier a005056C00008. Containers and services can be managed by stating the respective resource identifier after the command. Before starting containers/services, the respective item must be added to the platform, i.e., an add command for a container image or a service artifact must be specified. The items are stated in terms of URIs, at the moment usually files on the local file system.

For illustration, we present now a one typical sequence of CLI commands. We assume that in the example installation environment a005056C00008 is a listed resource and http://localhost/container.tgz is a valid container artifact with name myContainer as indicated by the associated deployment descriptor. Container myContainer has the service manager installed and will executed subsequent services. Moreover, we assume that http://localhost/simpleMesh.jar is a valid service artifact with name simpleMesh as defined by the included service descriptor. One sequence could be:

resources
  list
  ..
containers a005056C00008
  add http://localhost/container.tgz
  list
  start myContainer
  ..
services a005056C00008
  add http://localhost/simpleMesh.jar
  listArtifacts
  startAll simpleMesh
  listServices
  stopAll simpleMesh
  listServices
  remove simpleMesh
  ..
containers a005056C00008
  stop myContainer
  undeploy myContainer
  list
  ..
exit

Please note that for accessing the containers or the services of a certain resource, we need the respective resource identifier. Similarly, for starting and stopping containers or services. When exiting the CLI, the CLI container will end. The platform container will continue operating until you stop it explicitly, e.g., using docker ps to obtain the container identifier of the running container and docker stop <id> to ultimately stop the container.

This was a rather simple scenario involving one combined server and compute node. The more nodes are involved, the less this manual approach is feasible. For this purpose, the platform supports so called deployment plans, which assign services from artifacts to resources. A deployment plan can be set into force with the command deploy <path/URI> and stopped with undeploy <path/URI>. These are also the basic management operations of the platform user interface. You will find more information on the deployment plans in the documentation of the CLI.

Further information

For further information on using the individual parts, please consult the platform handbook linked on the main page. Please see also