Hey again fellow Flinkers, I shared [a post](https://stackoverflow.com/questions/77091173/potential-causes-for-invalidpidmappingexception-in-transactional-flink-job) a few days ago related to some transactional issues I’ve been trying to reproduce. In a nutshell, the job uses exactly once semantics for the sinks it writes to and we’ve seen an issue after adding a rebalance call to the job that’s been causing these InvalidPidMappingExceptions. The job will be in a running state, take a savepoint prior to the deployment, and restore from the savepoint (using the new version with the rebalance) and start throwing the exception. I can’t seem to consistently reproduce it, but it’s the only thing that changed between deployments of the job. Any recommendations to reproduce or on a potential cause would be greatly appreciated. There’s more detail in the post and I’d be happy to provide any level of detail as would be helpful.