Comments on: Loading a MySQL database in memory (with some Ruby help)

By: John Doyle

John Doyle — Sat, 13 Jun 2009 10:03:32 +0000

Many thanks just what I was looking for, innodb_buffer_pool_size large it is.

By: urbanus

urbanus — Thu, 21 Aug 2008 03:04:54 +0000

@gari <– COMMENT SPAM

Note the lame comment and the website link.

Do the ‘Net a favour and purge stupid posts like that, thanks.

By: gari

gari — Tue, 28 Aug 2007 19:24:29 +0000

Great site!0

By: Jeremy Cole

Jeremy Cole — Fri, 15 Jun 2007 17:58:09 +0000

If you have enough memory to hold the entire database in memory, why not just make innodb_buffer_pool_size large enough to do so. The speed should be the same (or similar) as moving the tables themselves into the MEMORY engine, without the data loss problems.

In my experience, MEMORY hasn’t been all that useful.

By: Jorge Bernal

Jorge Bernal — Thu, 14 Jun 2007 16:16:15 +0000

Wow, that’s easy! Thanks

By: Giuseppe Maxia

Giuseppe Maxia — Thu, 14 Jun 2007 16:09:33 +0000

A RAM disk on Mac OS X is quite easy, actually:

http://osxdaily.com/2007/03/23/create-a-ram-disk-in-mac-os-x/

Cheers

Giuseppe

By: Jorge Bernal

Jorge Bernal — Thu, 14 Jun 2007 15:35:38 +0000

I thought about a ramdisk for this, but I’m using OSX and I don’t know how to do it by now

By: Giuseppe Maxia

Giuseppe Maxia — Thu, 14 Jun 2007 14:54:37 +0000

Jorge,

instead of changing the engine type, there is an easier way.
Move the data directory to an in-memory filesystem.
In Linux, it’s usually in /dev/shm.

1. stop the server

2. cd /usr/local/mysql
3. mv data data1
4. cp -R data1 /dev/shm/data
5. ln -s /dev/shm/data /usr/local/mysql/data

6. start the server

Now the data is in memory, without any limitation.
(Warning: the data is in memory only! If your server shuts down, you lose all)

When you are finished, copy the data from the in-memory filesystem to the hard disk.

Cheers

Giuseppe