Formatting

City College of San Francisco - CS260A
Unix/Linux System Administration
Module: Filesystems I

Formatting

As you should have read when you researched the Internet, the term formatting has two meanings:

low-level formatting of modern disk drives is usually only done at the factory. At this time, the disk is burned-in, bad sectors and tracks are spared, and the low-level marks on the disk, which are used by the hardware later to verify offsets, are laid down. After this, if a modern drive develops bad sectors or tracks, the data on it should be backed up and the drive replaced. Low-level formatting is done on the drive as a whole.

The only type of media that should commonly be low-level formatted are floppies. Most systems have a formatting program for floppies. On linux it is fdformat

high-level formatting or erasing involves writing an empty filesystem on the device after it has been low-level formatted and partitioned. This operation is done on a partition by partition basis. The common term for high-level formatting on a Unix or linux system is creating a filesystem. We will learn about filesystem creation later. Note that although this operation is called erasing, the contents of the partition are not actually erased, and data can often be partially or completely extracted from these 'erased' partitions using modern data recovery techniques. Sensitive data should be more completely destroyed by using a tool to overwrite all the blocks on the disk.

Disk layout

The storage of data bits on a modern disk is extremely complicated, and beyond the scope of this class. Traditionally, however, the ideas were much simpler: storage (and location on the disk) was expressed in units of cylinders, sectors and blocks. Although you will from time to time see references to "cylinders" or "sectors", the only real measurement that has meaning today is block. When referring to physical disk blocks, the traditional size was 1/2 kb, or 512 bytes, but the context in which the term is used is important, as described below:

The block issue

The term block is used in several different contexts on a Unix system. In the discussion above, the block is the smallest indivisible unit of disk allocation, and it is usually 512 bytes. When the filesystem is created, filesystem space is also measured in blocks, where the block is the smallest amount of space that can be allocated on the filesystem. There the size of a 'block' is typically some number of kB that is a power of 2: 1kB, 2kB, 4kB or 8kB. In some other contexts, as we will see, the term block is used to refer to disk units that are 1kB in size. Due to this ambiguity, be very careful to determine the meaning of a block when it is being used as a measurement.

Preview question: Device files are used to access disk partitions. Log on to one of our classroom linux systems and check which devices are used to access filesystems using the df command. Physical devices will have a path that starts with /dev and does not have lv or vg in the name.

This page was made entirely with free software on linux:
Kompozer and Openoffice.org