If a metadata scheduled task which creates extraction records fails during processing, currently the start id for the next run will already have been updated, this is incorrect, we should try the task again from the same startid, the startid should not be incremented until successful processing is conducted.
If a metadata scheduled task which creates extraction records fails during processing, currently the start id for the next run will already have been updated, this is incorrect, we should try the task again from the same startid, the startid should not be incremented until successful processing is conducted.