Thursday, January 7, 2010

Refactoring team result

The team was busy testing available solutions for current situation. It was narrowed to, MySQL and Postgres because of relational nature of our data.

MySQL was faster and acting OK. But for something like 51M records it was harder to make it work properly and it request much more resource than Postgres.

Insert time was the same for them but Indexing time for MySQL grows improportional to the number of records.

Also the way MySQL create the index in lock situation make it a little unsuitable for our use.

Postgres seems to use much less resources than MySQL that makes it a better choice for our shares server environment.