Commit and Online Redo Log
Hello,
If we have got a process that makes inserts to a table of many many rows, for instance millions of rows, what considerations one should take in order to make a decision at what portions to issue commit?
"The online redo log is a set of two or more online redo log files that record all changes made to the database, including *uncommitted* and committed changes" - this sentence tells us that my changes to the database (in this particular case these changes are inserts) are going to be recorded on the redo log files.
But what if the free space on these files "ends" and at the same time I still have not issued commit?
My question is totally theoretical. I am not going to do these inserts since there many different efficient methods of collecting large volumes of data.
I just would like to know what would happen if the size of the files appear to be too "small" for such transaction? And in what way one could calculate the upper boundary providing that one knows the size in KB of each row (at least approximately)?
Or the whole process is totally different logically?
Please, explain!