We will add a simple action that runs a local shell command. Run the following from your docker host.
mkdir -p packs.dev/tutorial/actions
cp tutorial/actions/hello.yaml packs.dev/tutorial/actions
Get a bash shell in the stackstorm
container:
docker-compose exec stackstorm /bin/bash
Load the new action into StackStorm. Whenever you change the yaml file, you need
to run st2ctl reload
. Within the container, run the following:
root@aff39eda0bdd:/# st2ctl reload --register-all
... output trimmed ...
Now, let's run the action:
root@aff39eda0bdd:/# st2 run tutorial.hello
.
id: 58f67dbf33a99300bdc4d618
status: succeeded
parameters: None
result:
failed: false
return_code: 0
stderr: ''
stdout: Hello human!
succeeded: true
The action takes a single parameter name
, which as we can see above,
defaults to 'human' if name
is not specified. If we specify a value for
name
, then as expected, the value is found in result.stdout
.
root@aff39eda0bdd:/# st2 run tutorial.hello name=Stanley
.
id: 58f67dc533a99300bdc4d61b
status: succeeded
parameters:
name: Stanley
result:
failed: false
return_code: 0
stderr: ''
stdout: Hello Stanley!
succeeded: true
Congratulations, you have created your first simple action!
If you want to take advantage of concurrency, use a slight variation on the above. On the host, run:
mkdir -p packs.dev/tutorial/policies
cp tutorial/actions/hello-concurrency.yaml packs.dev/tutorial/actions
cp tutorial/policies/hello-concurrency.yaml packs.dev/tutorial/policies
Inside the stackstorm
container, run:
st2ctl reload --register-all
Open two terminals to the stackstorm
container. In the first, type (but don't execute):
st2 run tutorial.hello-concurrency name=1
In the second, type:
st2 run tutorial.hello-concurrency name=2
Now, execute the command in the first terminal, wait 5 seconds and then execute the command in the second terminal. After a second or so, you should see the following in the second terminal:
root@258b11849aa7:/# st2 run examples.hello-concurrency name=2
.
id: 590cec228964ad01567f61e3
status: delayed
parameters:
name: 2
result: None
If you run st2 execution list
before 10 seconds have elapsed, the status of the second action should
be "delayed". Between 10 and 20 seconds, the status of the second action should be "running". After
20 seconds, the status of the second action should be "succeeded".
root@ffc8bc7909c6:/# st2 execution list
+--------------------------+----------------------------+--------------+-------------------------+-------------------------------+-------------------------------+
| id | action.ref | context.user | status | start_timestamp | end_timestamp |
+--------------------------+----------------------------+--------------+-------------------------+-------------------------------+-------------------------------+
:
(snip)
:
| 5a366f07a1d7aa00ecfd3cef | tutorial.hello-concurrency | st2admin | succeeded (11s elapsed) | Sun, 17 Dec 2017 13:20:07 UTC | Sun, 17 Dec 2017 13:20:18 UTC |
| 5a366f0aa1d7aa00ecfd3cf2 | tutorial.hello-concurrency | st2admin | succeeded (18s elapsed) | Sun, 17 Dec 2017 13:20:10 UTC | Sun, 17 Dec 2017 13:20:28 UTC |
+--------------------------+----------------------------+--------------+-------------------------+-------------------------------+-------------------------------+
To perform a very basic end-to-end test of StackStorm, let's create a simple rule. Run the following from your docker host.
mkdir packs.dev/tutorial/rules
cp tutorial/rules/monitor_file.yaml packs.dev/tutorial/rules
Take a look at monitor_file.yaml
. The core.local
action is triggered when the
contents of /tmp/watcher.log
change.
Use docker-compose exec
to connect to the stackstorm
container:
docker-compose exec stackstorm /bin/bash
Run the following:
st2ctl reload
When we append to /tmp/watcher.log
, the sensor will inject a trigger and the
action will be executed. Now let's append a line to the file in the container.
echo "hello" >> /tmp/watcher.log
You should see that the action was fired:
root@ffc8bc7909c6:/# st2 execution list
+--------------------------+----------------------------+--------------+-------------------------+-------------------------------+-------------------------------+
| id | action.ref | context.user | status | start_timestamp | end_timestamp |
+--------------------------+----------------------------+--------------+-------------------------+-------------------------------+-------------------------------+
:
(snip)
:
| 5a36702fa1d7aa00373b785c | core.local | stanley | succeeded (0s elapsed) | Sun, 17 Dec 2017 13:25:03 UTC | Sun, 17 Dec 2017 13:25:03 UTC |
+--------------------------+----------------------------+--------------+-------------------------+-------------------------------+-------------------------------+
root@ffc8bc7909c6:/# st2 execution get 5a36702fa1d7aa00373b785c
id: 5a36702fa1d7aa00373b785c
status: succeeded (0s elapsed)
parameters:
cmd: 'echo "{''file_name'': ''watcher.log'', ''line'': u''hello'', ''file_path'': ''/tmp/watcher.log''}"'
result:
failed: false
return_code: 0
stderr: ''
stdout: '{''file_name'': ''watcher.log'', ''line'': u''hello'', ''file_path'': ''/tmp/watcher.log''}'
succeeded: true
Congratulations, you have created your first rule!
As an example of how to create a new action, let's add a new action called echo_action
.
First, on the host, we create the metadata file ./packs.dev/tutorial/actions/my_echo_action.yaml
:
---
name: "echo_action"
runner_type: "python-script"
description: "Print message to standard output."
enabled: true
entry_point: "my_echo_action.py"
parameters:
message:
type: "string"
description: "Message to print."
required: true
position: 0
Then, add the action script at ./packs.dev/tutorial/actions/my_echo_action.py
.
import sys
from st2common.runners.base_action import Action
class MyEchoAction(Action):
def run(self, message):
print(message)
if message == 'working':
return (True, message)
return (False, message)
When you rename, or create a new action, you must run st2ctl reload
inside the st2
container. Next, to initialize the virtualenv, run:
st2 run packs.setup_virtualenv packs=tutorial
Then you can run your action using the following:
st2 run tutorial.echo_action message=working
You should see output similar to:
.
id: 58c0abcff4aa45009f42dca3
status: succeeded
parameters:
message: working
result:
exit_code: 0
result: working
stderr: ''
stdout: 'working
'
Congratulations! You have successfully added your first action!
To add a simple mistral workflow, run the following from your docker host.
mkdir -p packs.dev/tutorial/actions/workflows
cp -R tutorial/actions/mistral-basic.yaml packs.dev/tutorial/actions/mistral-basic.yaml
cp -R tutorial/actions/workflows/mistral-basic.yaml packs.dev/tutorial/actions/workflows/mistral-basic.yaml
Use docker-compose exec
to connect to the stackstorm
container:
docker-compose exec stackstorm /bin/bash
Within the container, run the following:
st2 action create /opt/stackstorm/packs.dev/tutorial/actions/mistral-basic.yaml
st2 run examples.mistral-basic cmd=date -a
The st2 run
command should complete successfully. Please see
mistral documentation
for more details about this basic workflow.
Congratulations, you have created your first mistral workflow!