1 master node+1 delayed backup node
At this time, the delayed backup node is configured to be delayed for 12 hours relative to the master node, then the 12-hour data (oplog) is
Will pull 12 hours of data to the delay node in advance, and then delay the application of oplog?
Or does it only exist in the master node, that is, the oplog is pulled with delay?
In case 2 above, should the oplog size of the primary node be set large enough (enough to guarantee at least 12 hours of oplog storage)?
If the delay is longer (1~2 days), then the corresponding oplog needs to be adjusted more?
By default, I remember 500MB, right? Only one point is required, and this capacity is sufficient to afford the delayed amount of data.
This can be estimated based on the output of db.stat (). there is an average capacity of object. based on that, you can estimate the amount of data you may need for a delay of 12 hours.