In short, is there a way to use the Deployment Automation Manager to run a flyway repair command to recover from failed DDL executed via Flyway, or is there anything planned for a future version of the plugin to allow this?
Some more background:
I have been through the documentation of the Automated Deployment Manager here: https://community.appian.com/w/the-appian-playbook/198/deployment-automation
Using what is documented, I have been successful in using the `-flyway_path` and `-ddl_ds` options to deploy database changes to an appian instance hosted by appian cloud, against the underlying MariaDB.
In the process of proving the above, I learnt that MariaDB does not rollback DDL when it rolls back database transactions. As a consequence flyway leaves the flyway_schema_history in a state that requires manual intervention and that blocks further flyway migrations until a repair is done.
Flyway has a documented way from recovering from failed migrations, which is a command for repairing the flyway_schema_history table. Documented here: https://flywaydb.org/documentation/command/repair
Our preferred way of recovering from these situations is for scripts authors to apply the appropriate fixes to the flyway migrations and then repeat the deployment to roll the database forward, and recovering the flyway history should be implicit / automatic.
Discussion posts and replies are publicly visible