Skip to content

start_stop.md

Latest commit

 

History

History
89 lines (62 loc) · 3.35 KB

start_stop.md

File metadata and controls

89 lines (62 loc) · 3.35 KB

Advanced start, stop and status

Starting

All services

To spin up the entire monaca-docker (including both log and metric pipelines) one must execute:

$ docker-compose -f log-pipeline.yml -f docker-compose.yml up

from the root of the monasca-docker repository.

Invidual pipelines can still be launched. See below for more details

Metric pipeline

For metric pipeline it suffices to just execute

$ docker-compose up

or

$ docker-compose -f docker-compose.yml up

There is a 3rd approach which requires passing all (some) of the members of metric pipeline as arguments after up. For example:

$ docker-compose up kafka zookeeper monasca-notification

It will result in launching only kafka, zookeeper and monasca-notification.

Log pipeline

The log pipeline is a bit more complicated. As it is an addition to the metric pipeline, some of the services are shared between each and thus defined in docker-compose.yml without a duplicate in log-pipeline.yml. That affects the way you could launch just the log pipeline. See the examples below to get a grip:

  • both pipelines docker-compose -f log-pipeline.yml -f docker-compose.yml up
  • just log pipeline docker-compose -f log-pipeline.yml -f docker-compose.yml up {{ log_pipeline_members }} where log_pipeline_members resolves to all the services defined in log-pipeline.yml file + some of the services that are shared between i.e kafka, zookeeper, keystone, mysql. Obviously all the -init containers corresponding to aforementioned services should be included in launch as well.

Invidual services

Launching individual services has been described above to some extent. In greater detail, launching an individual service is as easy as specifying what service(s) you wish to launch as an argument to the docker-compose up command. For example, to launch monasca-log-transformer you execute something like docker-compose -f log-pipeline up log-transformer. The tricky point is to know the "dependencies" of single service. These "dependencies" need to be available, either externally or launched together with the desired service.

Stopping

Stopping the services it similar to starting them. The difference is that instead of doing docker-compose up you do docker-compose stop. With this command you can:

  • stop everything docker-compose -f log-pipeline.yml -f docker-compose.yml stop
  • stop service(s) docker-compose -f log-pipeline.yml -f docker-compose.yml stop {{ services }}

An important thing to mention is that stop does not result in removing the containers from the system. In order to stop and execute a bit of cleaning replace stop with down.

Getting status

To know if the the service is up and running, you can execute docker-compose ps. That command is similar to plain docker ps. However it allows to use the easier to remember service name instead of auto-generated labels docker assigns to containers. For example:

$ docker-compose ps kafka
        Name             Command    State    Ports
----------------------------------------------------
monascadocker_kafka_1   /start.sh   Up      9092/tcp