I have 32GB memory. SSD and NonSSD drives.
I installed Linux to NonSSD drive but I prefer SSD for /var directory and I didn't create swap space because there is high memory.
I know physical memory + swap = virtual memory. I know there are hugepages and memory mapped files features of linux. Memory mapped files uses virtual memory and I don't know a strategy to process a file. Is there a benefit of swap even if there is high physical memory.
I'm going to process 3GB gzipped file to import mysql database. I'm going to create application cpp boost library gzip memory mapped stream and mysql native client library.
- My first target is detecting diff of files. I'm going to use
ON DUPLICATE KEY UPDATE attempt_count = attempt_count + 1.
- Also I need fulltext search feature and I need to filter some criteria like string length.
- And I should integrate a management interface.
What is the best configuration for this?