In other cases, when Linux must work alongside other already existing partitions, you need to choose manual partitioning.
4.2.13.1. Guided partitioning
The guided partitioning tool offers three partitioning methods, which correspond to different usages.
Figure 4.9. Guided partitioning
The first method is called “All in one partition”. The entire Linux system tree is stored in a single filesystem, corresponding to the root / directory. This simple and robust partitioning fits perfectly for personal or single-user systems. In fact, two partitions will be created: the first will house the complete system, the second the virtual memory (swap).
The second method, “Separate /home/ partition”, is similar, but splits the file hierarchy in two: one partition contains the Linux system (/), and the second contains “home directories” (meaning user data, in files and subdirectories available under /home/).
The last partitioning method, called “Separate /home, /usr, /var, and /tmp partitions”, is appropriate for servers and multi-user systems. It divides the file tree into many partitions: in addition to the root (/) and user accounts (/home/) partitions, it also has partitions for applications (/usr/), server software data (/var/, and temporary files (/tmp/). These divisions have several advantages. Users can not lock up the server by consuming all available hard drive space (they can only fill up /tmp/ and /home/). The daemon data (especially logs) can no longer clog up the rest of the system.
BACK TO BASICS Choosing a filesystem
A filesystem defines the way in which data is organized on the hard drive. Each existing filesystem has its merits and limitations. Some are more robust, others more effective: if you know your needs well, choosing the most appropriate filesystem is possible. Various comparisons have already been made; it seems that ReiserFS is particularly efficient for reading many small files; XFS, in turn, works faster with large files. Ext3, the default filesystem for Debian, is a good compromise, based on the two previous versions of filesystems historically used in Linux (ext and ext2). You could also choose its successor ext4, which overcomes certain limitations of ext3 and is particularly appropriate for very large capacity hard drives. If you are particularly brave, you could experiment with the very promising btrfs, which includes numerous features that require, to this day, the use of LVM and/or RAID.
A journalized filesystem (such as ext3, ext4, btrfs, reiserfs, or xfs) takes special measures to make it possible to return to a prior consistent state after an abrupt interruption without completely analyzing the entire disk (as was the case with the ext2 system). This functionality is carried out by filling in a journal that describes the operations to conduct prior to actually executing them. If an operation is interrupted, it will be possible to “replay” it from the journal. Conversely, if an interruption occurs during an update of the journal, the last requested change is simply ignored; the data being written could be lost, but since the data on the disk has not changed, they have remained coherent. This is nothing more nor less than a transactional mechanism applied to the filesystem.
After choosing the type of partition, the software calculates a suggestion, and describes it on the screen; the user can then modify it if needed. You can, in particular, choose another filesystem if the standard choice (ext3) isn't appropriate. In most cases, however, the proposed partitioning is reasonable and it can be accepted by selecting the “Finish partitioning and write changes to disk” entry.
Figure 4.10. Validating partitioning
4.2.13.2. Manual Partitioning
Manual partitioning allows greater flexibility, allowing the user to choose the purpose and size of each partition. Furthermore, this mode is unavoidable if you wish to use software RAID.
IN PRACTICE Shrinking a Windows partition.
To install Debian alongside an existing operating system (Windows or other), you must have some available hard drive space that is not being used by the other system in order to be able to create the partitions dedicated to Debian. In most cases, this means shrinking a Windows partition and reusing the freed space.
The Debian installer allows this operation when using the manual mode for partitioning. You only need to choose the Windows partition and enter its new size (this works the same with both FAT and NTFS partitions).
The first screen displays the available disks, their partitions, and any possible free space that has not yet been partitioned. You can select each displayed element; pressing the Enter key then gives a list of possible actions.
You can erase all partitions on a disk by selecting it.
When selecting free space on a disk, you can manually create a new partition. You can also do this with guided partitioning, which is an interesting solution for a disk that already contains another operating system, but which you may wish to partition for Linux in a standard manner. See the previous section for more details on guided partitioning.
BACK TO BASICS Mount point
The mount point is the directory tree that will house the contents of the filesystem on the selected partition. Thus, a partition mounted at /home/ is traditionally intended to contain user data.
When this directory is named “/”, it is known as the root of the file tree, and therefore the root of the partition that will actually host the Debian system.
BACK TO BASICS Virtual memory, swap
Virtual memory allows the Linux kernel, when lacking sufficient memory (RAM), to free a bit of storage by storing the parts of the RAM that have been inactive for some time on the swap partition of the hard disk.
To simulate the additional memory, Windows uses a swap file that is directly contained in a filesystem. Conversely, Linux uses a partition dedicated to this purpose, hence the term “swap partition”.
When choosing a partition, you can indicate the manner in which you are going to use it:
format it and include it in the file tree by choosing a mount point;
use it as a swap partition;
make it into a “physical volume for encryption” (to protect the confidentiality of data on certain partitions, see below);
make it a “physical volume for LVM” (this concept is discussed in greater detail later in this chapter);
use it as a RAID device (see later in this chapter);
or the choice not to use it, and therefore leave it unchanged.
4.2.13.3. Configuring Multidisk Devices (Software RAID)
Some types of RAID allow the duplication of information stored on hard drives to prevent data loss in the event of a hardware problem affecting one of them. Level 1 RAID keeps a simple, identical copy (mirror) of a hard drive on another drive, while level 5 RAID splits redundant data over several disks, thus allowing the complete reconstruction of a failing drive.
We will only describe level 1 RAID, which is the simplest to implement. The first step involves creating two partitions of identical size located on two different hard drives, and to label them “physical volume for RAID”.