IT 4100 : File Systems and Storage Technologies

Windows-based filesystems

Preliminary Terms

• Files are comprised of groups of clusters
• Clusters are made up of groups of sectors
• So, a cluster is the smallest unit of disk space that can be allocated to a file, also called allocation units.

Partitions

• Still exist
• Chop your disk up into sectors that can participate in making clusters
• Sectors are uniform size per physical disk
  • Standard size is 512
  • sudo fdisk -l /dev/sda5 (view sector size)
  • but if your sector size is different, you would have to pass-in a corresponding sector size.
  • Newer hard drives can write 4096 sectors
• A cluster is a group of sectors.

FAT

• Most simplistic FS supported by WindowsNT.
• Consist of a file allocation table
• Still found on small devices
  • USB, flash memory

FAT

A disk formatted with FAT is allocated in clusters, whose size are determined by the size of the volume. When a file is created, an entry is created in the directory and the first cluster number containing data is established. This entry in the FAT table either indicates that this is the last cluster of the file, or points to the next cluster.

FAT Disk layout

• Boot block
• File allocation table
  • And a backup or duplicate table
• root directory
• file data area

FAT boot block (AKA partition boot sector)

• Can hold special bootstrap program for loading OS into memory.
• MBR can use this area
• See this

The File allocation Table

• The FAT occupies one or more blocks immediately following the boot block.
• Can have more than one (at least 2 should be kept)
• For 16-bit fat, each entry in FAT is 2 bytes in length.
• there is one entry in the FAT for every cluster (data area block) on the disk. So,
  • Making clusters very large results in a small FAT, but also wastes space as no two files can share the same cluster
  • Making clusters very small results in a very large FAT, which also wastes time, but more importantly, also degrades performance.

The File allocation Table

• Usually contained in the first sector of cluster 2. (Generally 2 copies are kept)
• A fat entry contains:
  • A pointer to the next cluster that should be accessed
  • Files that use more than a single cluster need to maintain where the next portion of the file can be found.
• See this image.

The root directory

The Root Directory, sometimes referred to as the Root Folder, contains an entry for each file and directory stored in the file system. This information includes the file name, starting cluster number, and file size. This information is changed whenever a file is created or subsequently modified. (Also includes the create time, last access, etc… see here)

The Data Area

The Boot Record, FATs, and Root Directory are collectively referred to as the System Area. The remaining space on the logical drive is called the Data Area, which is where files are actually stored. It should be noted that when a file is deleted by the operating system, the data stored in the Data Area remains intact until it is overwritten.

FAT

• Writing to FAT tabled is time consuming
• No organization to FAT directory structure, files are given first open location on the drive

Fat

File slack

The difference between the files logical size and the size of the cluster. OS doesn’t write there, so could store old data or secret data.

Example

• Create 100 MB partition
• sudo mkfs.vfat -s 16 /dev/sda5 (man mkfs.vfat)
• check it out
  • sudo fsck.vfat -vr /dev/sda5
  • Assuming you have a file on this partition, this command will also tell you how many bytes per sector, sectors per cluster, and how many clusters the file takes up.
  • sudo file -s /dev/sda5

Example

The above command utilized 16 sectors per cluster. If each sector is 512 Bytes, we could calculate that each cluster is then 8192 bytes in size. If you then create a file, you should be able to determine how many clusters it will use and then verify it using the du command.

Example

This example utilized 2 sectors per cluster (i.e. mkfs.vfat -s 2 -F 32 /dev/sda25). 14493 is the size of the contents of the file (i.e. fallocate -l 14493 foo.txt). ¹⁴⁴⁹³⁄₁₀₂₄ = 14.15 (round to 15)

    joe@12g:~/2sectorpercluster$ ls -l z.txt 
    -rwxrwxr-x 1 joe joe 14493 Apr  3 15:04 z.txt
    joe@12g:~/2sectorpercluster$ du z.txt 
    15  z.txt

Example

Create a file that is one less than a multiple of the cluster size… For example if you have 2 sectors per cluster, 16 clusters would be a file size of 15360. What if you create a file of 15359? What if you create a file of 15361? See the output of the du command.

Advantages

• Files CAN be undeleted

Disadvantages

Preferably, when using drives or partitions of over 200 MB the FAT file system should not be used. This is because as the size of the volume increases, performance with FAT will quickly decrease. It is not possible to set permissions on files that are FAT partitions.

MISC

• FAT16 contains 2 bytes per cluster within the FAT table
  • 2GB max file size
• FAT32 contains 4 bytes per cluster within the FAT table
  • 4G max file size

Today, FAT file systems are still commonly found on floppy disks, USB sticks, flash and other solid-state memory cards and modules, and many portable and embedded devices.