Attribute on Task to allow Workflow to continue if there is a failure

[#1]

I have this one task in a workflow that is failing, but it is not an important task, and I would like to see the workflow “keep chugging”. Let me confirm that there is no way to do that? It would be nice if we had, on the task itself, the ability to have a flag or checkbox that indicates “continue running workflow” in the case of an error, so that non-critical tasks don’t bring down the whole enchilada.

Maybe the answer is to break the workflows into a series of workflows (I have considered that). But even then, the VM will come up with a yellow exclamation if that non-critical task fails - which, I suppose, would be okay as long as the user could “acknowledge” it and set the VM back to green arrow. But I think once a VM comes up with a yellow exclamation, it sits that way indefinitely right? Or, does it flip back to green after some length of time?

[#2]

Actually, the -w option on the bash script did not seem to work, after testing. Not sure if Morpheus scripts can handle option parsing or not. I had to back out the -w idea.

[#3]

I did exactly that on this bash script. I am prompting for an LDAP group, and after I check validation on the group, I add that group to sudoers. But, I didn’t want the whole workflow falling down if someone fat fingered the LDAP group (we can always re-run the task and correct that). So what I did, was add getopt to the bash script, and added a -w option. If the -w option is passed in, it will bark warnings but not outright fail with a non zero exit, allowing the more important tasks in the workflow to continue. There are indeed “some” tasks that I consider fatal to the workflow, and some that I would want yellow as a “look into” warning, while others, I might just print out informational stuff (even warnings) but leave things green.

[cbunge]

Hi Mark!

We’re actually working on adjustments to the automation engine to give better control to the users on success failures and actions that should follow. There’s currently no ETA on that.

In the mean time, I believe the yellow ! goes away after a while but not exactly sure on the on the interval but I think it’s controlled by the monitoring frequency Morpheus creates for each provisioning.

A side note, if you want to fail forward if you can capture the output of the script and force exit 0 it will move forward no matter what. So you can write the errors on the task and still exit currently.