SDDMS V4 API doesn’t implement lockid
The V3 API implements a lockid (sent in the x-seismic-dms-lockid HTTP header) when registering a new dataset, a ‘lock’ endpoint which specifies the openmode for a dataset when reading/writing/deleting it and a ‘close’ endpoint that releases the lock.
That ensured that it wasn't possible to read the dataset unless it’s completed writing and is crucial when running jobs that takes hours (or even days) that are creating the dataset and that it’s not possible to delete the dataset while it is being read/written. A dataset should not ideally not be registered in the catalog until the write lock is released indicating the dataset was written successfully. However, lockid isn't implemented in V4. How can we implement proper error handling if an error is made during the creation of the dataset?