Tasks stop when conductor server dies

[AdityaMishraGit]

Hi, We are using netflix conductor with postgres database , but the tasks are getting stuck if the conductor server dies. We have a requirement where the task should auto start once we restart the conductor server. Is there a way to do so or are we missing something?

[v1r3n]

Hi @AdityaMishraGit do the task remain IN_PROGRESS status or SCHEDULED status after the workflow restarts?

[AdityaMishraGit]

Hi , thanks for replying. It remains on SCHEDULED state

[v1r3n]

Can you check the workflow sweeper is running? It is enabled by default unless configured to disable. The following property controls it

conductor.workflow-reconciler.enabled

[AdityaMishraGit]

The property is enabled . I was able to repair the workflows stuck by the verifyAndRepair api end point , thanks to the hint.

I guess I will need to write a schedular to repair my stuck workflows

[aravindanr]

fwiw, enabling the property conductor.workflow-repair-service.enabled runs the verifyAndRepair on a schedule.