Description | Data can sneak in on other channels during an initial load and cause issues or errors. This can happen because each channel is given a chance to extract data during an extract. If changes come in for other channels during a reload the data on those channels can be sent before the reload is done. If we do the load on the channels they are suppose to be on, then this won't happen. I guess another option that I had not thought of would be to only extract from the reload channel if it has events assigned to it. We wouldn't send other channels until the reload channel has cleared. I'd like to add column(s) to outgoing_batch to make it easy to identify reload batches. I was thinking of adding columns for each event_type. The column would have the count of events of each type in the batch. The columns would be: * reload_count * insert_count * update_count * delete_count * other_count |
---|