Can I setup capped collection on machine A and non-capped replication of that collection on machine B (so that machine B has all the data that was written on machine A)?
The reason behind:
The data on machine A fills in very quickly (over 10GB per day) and I want to make sure that this data does not exhausts the disk space and brings down the whole server.
This question was recently asked and answered on the MongoDB Users list so I'll just link to the answer there but for locality, I'll summarize:
"No, there is no option for that. The goal of replication it have the same data on all nodes."
The answer there also includes a possible way to do this, but it would be completely unsupported (and unadvised).