Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have a Fedora 17 system containing three (3) hard disks [ /dev/sda, /dev/sdb and /dev/sdc ].

The /dev/sda disk is the "boot drive" for my system. I have set the system up to automount /dev/sdb and /devsdc. I have a database application [MySQL] that I would like to have installed only on /dev/sdb, but I do not know how to accomplish this. [I intend to use /dev/sdc as the location for system backups.]

Would someone please tell me how to accomplish this? Under Windows 7 (for example) this task is trivial, but I have never seen any documentation for Linux. Thank you in advance for your help.

share|improve this question
Linux isn't Windows, and so you are going to have a lot of (bad?) habits to unlearn. Like this one. The first thing to learn is how a UNIX/Linux unified filesystem works. – Michael Hampton Nov 10 '12 at 19:14

Not to be too blunt, but everything you know about disk storage is wrong!
(Or at least based on a fundamentally broken paradigm that Windows inherited from DOS, which inherited it from CP/M).

A brief and woefully incomplete introduction to Unix filesystems

At the user/filesystem level Unix doesn't care what physical device it's using. There is only the one true great and bountiful directory hierarchy (tree), rooted at / (root), under which all things live.

Filesystems are created on disk devices (/dev/sda1, /dev/sda2, /dev/sdb1, etc.) and mounted at the location where you want to logically access them.

For example, /dev/sda1 (Windows drive C:, the first hard disk) is probably mounted at / (root).
You may then have /dev/sda2 mounted at /usr containing all of your "userland" programs, and /dev/sda3 mounted at /home containing all the user home directories. /dev/sdb1 might be mounted on /var to hold log files so they don't fill up your boot drive, etc.

The physical structure (which disks/partitions are associated with which mount points) is something you decide on when building the system, based on projected need/growth.
The logical structure (directory names and what goes in them) is standardized, and documented in your system manual (the hier man page -- hier(7) from a Linux system, and from a BSD system).

You can see which partitions (filesystems) are mounted where by running the mount command:

[mgraziano@monitor ~]$ mount
/dev/da0s1a on / (ufs, local)
devfs on /dev (devfs, local)
/dev/da0s1h on /home (ufs, local, soft-updates)
/dev/da0s2d on /opt (ufs, local, soft-updates)
/dev/da0s1g on /tmp (ufs, local, soft-updates)
/dev/da0s1e on /usr (ufs, local, soft-updates)
/dev/da0s1f on /usr/local (ufs, local, soft-updates)
/dev/da1s1d on /usr/local/pgsql/data (ufs, local, soft-updates)
/dev/da3s1d on /usr/local/pgsql/data/pg_xlog (ufs, local, soft-updates)
/dev/da2s1d on /usr/local/pgsql/data/xlog_archive (ufs, local, soft-updates)
/dev/da0s1d on /var (ufs, local, soft-updates)

and you can get more information (like space available) by using the df command:

[mgraziano@monitor ~]$ df -h
Filesystem     Size    Used   Avail Capacity  Mounted on
/dev/da0s1a      2G     73M    1.7G     4%    /
devfs          1.0k    1.0k      0B   100%    /dev
/dev/da0s1h     18G    6.3M     16G     0%    /home
/dev/da0s2d     63G    7.4M     58G     0%    /opt
/dev/da0s1g      2G     14k    1.8G     0%    /tmp
/dev/da0s1e    9.7G    810M    8.1G     9%    /usr
/dev/da0s1f     14G    1.9G     11G    14%    /usr/local
/dev/da1s1d    266G     21G    223G     9%    /usr/local/pgsql/data
/dev/da3s1d    130G    128M    120G     0%    /usr/local/pgsql/data/pg_xlog
/dev/da2s1d    266G    2.1G    242G     1%    /usr/local/pgsql/data/xlog_archive
/dev/da0s1d    9.7G    308M    8.6G     3%    /var

The results above are from a FreeBSD system, so the device names are different, but the general format of the output is the same.

So what's the answer to your question? -- "You're thinking about it wrong."

When you install a package like MySQL on a Unix system it will (hopefully) adhere to the conventions in the hier documentation. For example:

  • MySQL's configuration files will probably go in /etc/mysql/
  • The MySQL database files will probably go in /var/db/mysql/
  • The MySQL client binary will probably be installed as /usr/bin/mysql
  • The MySQL dump program will probably be installed as /usr/sbin/mysqldump
  • The MySQL daemon (the DB server program) will probably be /usr/sbin/mysqld
  • Startup scripts will probably be installed in /etc so the DB starts when the system boots.

What matters for the installation (and the user) is the logical location in the filesystem hierarchy. The concept of the underlying physical disk is abstracted away entirely (except if you run out of space, and the assumption is that you (the system administrator) have made your partitions large enough to accommodate what's being installed).

The abstraction offered by Unix filesystems has a number of benefits, among them the ability to cover mistakes made in the initial partitioning.
For example if the /var partition on your system is on is only 10GB but you expect MySQL to store 100GB of data you might go out and buy a new hard drive, install it, and mount it at /var/db/mysql so you have enough room for your database.
As an example, look at the df output I provided above - /usr/local/pgsql/data (my Postgres database directory) resides on a separate, larger partition from /usr/local (which itself is separated from /usr).

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.