I have a table with ~1.4 billion records. I want to check how many records are exactly there.
I have a Percona 5.5 server running on Ubuntu 12.04. The server has 80 GB of memory.
The table itself is ~100 GB.
I issue a simple command
Select count(*) from table
I then, monitor the INNODB engine. I get a reading of 35-45,000 reads/second
If my calculations are correct, it will need 9-18 hours just to do this simple command.
I thought that using databases would improve the speed of handling vast amounts of data. Is this realistic? Can I make my server faster?
I used the Percona my.cnf wizard to set all my variables.