Mutliple mysqld instances on same machine

Hey folks!

Two years ago i was reading this blog-post (…nces-of-mysql/ ) and decided to start separating mysql users to multiple instances.

Well, i really don’t know why i did that, i believe mostly as a secured enhancment?

Each customer has his own mysqld-process where all his databases are located.
To be honest, each user has only one database and didn’t have any knowledge of that.
So in total from 10 users (10 instances) each user database-setup is only 10 MB large, expect one community which is using 2 GB in size.

My question is now, do i have any advantages beside the little security thing or do i have more and more disadvantages?

The machine itself is a i7-6700 with 64GB DDR4 and RAID1 SSD setup.

per user setup, except community:

innodb_buffer_pool_dump_at_shutdown = ON
innodb_buffer_pool_load_at_startup = ON
innodb_buffer_pool_instances = 1
innodb_buffer_pool_size = 20M
innodb_file_per_table = 1
innodb_log_buffer_size = 16M
innodb_log_file_size = 2M
innodb_log_files_in_group = 2
innodb_page_size = 16K
innodb_thread_concurrency = 1
key_buffer_size = 16M
net_buffer_length = 16K
query_cache_size = 0
thread_cache_size = 4

same like per user setip with additional:

innodb_buffer_pool_instances = 8
innodb_buffer_pool_size = 8G
innodb_log_file_size = 625M
join_buffer_size = 2097152
table_open_cache = 3000

Can you give me a short briefing if this would be fit the needs of best performance or should i restore to one mysqld instance?
If yes, which settings would you prefer?

Thanks for helping to keep better performance.

For specific tuning questions, and more in-depth recommendations, I would suggest having a look at

It would indeed seem more secure to split different customers instances across different mysqld’s. While it may be possible to secure a single mysqld instance quite well, I expect there would still be loopholes.