Allow Workflow to save error and continue execution

[jeffgao]

Description:
We have post configuration workflow with about 20 tasks with some bash and python. The tasks include LVM resizing, repo config, OS firewall changes, email notification etc.

During 10 tasks of post configuration workflow

In some cases if one of the task failed, we want to save the error messages and state, continue next task. At the end of workflow, we can give all error message stacks.

For example our repository configuration failed, we know it’s firewall related, we save the error message, continue with LVM resizing.

The the end, we give message that VM sizing is good, but repository config failed. Can we have a way to stack message, continue task, and last task get all message out?

---

Example/Use case(s):
We have post configuration workflow with about 20 tasks with some bash and python. When we have none critical failure with task, we want to continue worklow.

With critical error, we want to mark the error, jump to email notification, send user the error message.

---

[cbunge]

This functionality already exists in 6.3.x and will be rolled into the 7.x LTS soon. On tasks there is an option to continue on error.

Also, if you take a look at this patching postrecently done it will show an example of pulling history of executions. Mine looks generally at Job executions but you can refer to the instance history and report back on status of tasks. Something like this could look retroactively at the workflow and group errors if desired.

[jeffgao]

We have 6.3.2, no “CONTINUE ONE ERROR” checkbox

Also is anyway for code to save task1 error msg, then task10 to retrieve error msg out?

[cbunge]

The output of all tasks is saved. The patching post I linked references the error block of a task, but could be specifically pulling the stdout.