Subject | Re: Forced write, page size and buffer size |
---|---|
Author | Dmitry Yemanov |
Post date | 2014-08-19T09:12:57Z |
18.08.2014 22:51, Aldo Caruso wrote:
cache size (2048 pages) but it's worth double checking.
Dmitry
>What is FB architecture: SS / CS / SC? I suspect SS given your default
> For reliability reasons, I decided to turn on forced writes on a
> database running on Linux.
cache size (2048 pages) but it's worth double checking.
> 1) Does it makes sense to activate forced writes on Linux ?It depends on what reliability means to you.
> 2) Is increasing the page size the right approach to improve performanceThe cost is a higher concurrency for the same pages under parallel load.
> ? Which are the drawbacks of setting page size to 16K ?
> 3) The same question stands for cache pages: is it useful ( or hasIt would surely not hurt.
> disadvantages ) to have so many cache pages as there are on disk pages (
> provided there is enough RAM size ) ?
Dmitry