Troubleshooting FAQ

This troubleshooting guide provides a list of frequently asked questions that point to common troubleshooting techniques or external Studio Conductor documents where necessary.

Deleting a deployment fails with the error message: Deployment still has active nodes

First uninstall the deployment (i.e. run the “uninstall” workflow on the deployment) and then run the delete on the deployment again.

A deployment install failed, how do I find out what went wrong?

By searching through the Studio Conductor logs and events you should be able to ascertain what went wrong with the install. The following Studio Conductor page here should detail how to retrieve events and logs for deployment workflows from the Conductor Management Console.

Where are the Studio Conductor service log files?

There are many different services running on the Conductor Manager most of which an operator should not need to access in day-to-day operation of the platform. However in the case where a Studio Conductor service cannot be restarted or appears to not be working these service log files are detailed here in this document.

What do the different status badges shown on the topology diagram in the Conductor Management Console mean?

The following document details these status badges here.

How do I know if a deployment is installed or not?

The status badges on each of the nodes in the deployments topology diagram indicate which nodes have been installed. See the following document for details of these status badges here.

Where can I see the inputs and outputs of a deployment?

You can see these in Conductor Management Console in deployment page (in Deployments page click on specific deployment to enter it).

How can I fix a failing deployment install?

If a deployment installation is failing with a runtime error (for example, the wrong credentials were provided), you should stop the installation and remove the deployment before you run the command again. To do that, run (replace DEPLOYMENT_NAME with an appropriate value):

cfy executions start stop -d DEPLOYMENT_NAME -p ignore_failure=true
cfy executions start uninstall -d DEPLOYMENT_NAME -p ignore_failure=true
cfy uninstall DEPLOYMENT_NAME

Fix the mistake and try again. If you run the uninstall commands above and get this error message:

An error occurred on the server: 404: Requested `Deployment` with ID `DEPLOYMENT_NAME` was not found

Just delete the “DEPLOYMENT_NAME” blueprint and try the install command again (read about blueprints and deployments commands).

How can I check to see if the Conductor Manager is operating correctly?

There is a way to find out this information using the Conductor CLI, this method is detailed in the “Troubleshooting techniques” section below.

We have a stuck deployment with ‘starting’ or ‘cancelling’ state?

Sometimes you will have deployments you can’t delete via the UI / CLI / API. These simple instructions will help you remove these.

First, make sure to cancel stuck execution:

cfy executions list -d <deployment id>

cfy executions cancel <stuck execution id>

From the manager machine:

psql -U Conductor -W Conductor -h localhost -d cloudify_db –p 5432 # -p 15432 for CFY HA cluster
pass: Conductor

To remove a stuck execution:
delete from executions where status='cancelling';

To exit the CLI:
\q

Once you are done – refresh the UI. You should see these execution removed from the list.

You should run uninstall to clean manager from deployment data:

cfy uninstall -d <deployment id>

What to do if an active manager is stopped or fails during workflow execution?

This will result in a stuck deployment. Follow the steps described in the previous answer.

What to do if an active manager is stopped or fails during plugin upload?

This will result in an inconsistent plugin management. Follow these steps:

cfy plugins delete <plugin id>

cfy plugins upload <plugin file path> -y plugin.yaml path

Troubleshooting techniques

How to run Conductor CLI commands

In order to run the Conductor CLI command ‘cfy’ you need to ensure that you have installed the Conductor CLI

rpm –qa | grep cloudify-cli

Make sure your CLI profile is properly set:

$ cfy profiles use <manager IP>

How to retrieve the status of the Conductor manager or cluster via the cli

How to retrieve the status of the Conductor Manager or cluster via the cli

$ [centos@cm-1 ~]$ cfy status
Retrieving manager services status... [ip=10.1.1.41]
Services:
+--------------------------------+--------+
|            service             | status |
+--------------------------------+--------+
| Cloudify Console               | Active |
| Manager Rest-Service           | Active |
| PostgreSQL                     | Active |
| AMQP-Postgres                  | Active |
| File Sync Service              | Active |
| RabbitMQ                       | Active |
| Webserver                      | Active |
| Cloudify Composer              | Active |
| Management Worker              | Active |
+--------------------------------+--------+


$ cfy cluster status
Retrieving Cloudify cluster status... [ip=10.1.1.41]
Current cluster status is OK:

Cluster status services:
+--------------------------------+----------+
|            service             |  status  |
+--------------------------------+----------+
| manager                        |    OK    |
| db                             |    OK    |
| broker                         |    OK    |
+--------------------------------+----------+