Deployment-IDs & undeploy

[dognose]

Hello,

currently we are searching for a way to undeploy a process Definition and remove all references.
The DeploymentManager along with the RepositoryManager seems to be a good point to start.

- First get all deployed Processes of a certain key.
- Then get the deployment id based on the actual process keys (key:xxx:yyy)
- Then undeploy this deployment, using cascade=true.

The only question I could not figure out so far: Are deployment ids *exclusive* for a certain process,
or could 2 different process definitions be part of the same deployment id?

i.e. if i deploy - let's say process1 and process2 within the same (physical) deployment -
are they assigned to different deploymentIDs, or do they share the same?

[sathish1]

There is one deployment and a new record will be created in the ACT_RE_DEPLOYMENT table and every time u start a process a new process instance will be created and those records are in the ACT_RE_PROCDEF. There is good explanation in the user-guide. <link> http://www.activiti.org/userguide/#N165AD </link>

[sahusunil]

Ideally deployment Id is exclusive to every process definition when you deploy it from activiti explorer but if you run a JUnit test to deploy the process definitions, it might allow you use to same deployment Id. This unique deployment id goes to ACT_RE_DEPLOYMENT table and ACT_RE_PROCDEF table which stores the process definition information has a link for this deployment Id. I think that's how undeploy logic works because it has reference of associated process definitions in ACT_RE_PROCDEF table. When you start the process from process definition, process instance info goes to ACT_RU_EXECUTION table (not ACT_RE_PROCDEF) and associated task entries goes to ACT_RE_TASK table.