How Much Disk Space Does a 1-Byte File Actually Occupy?

This article explores how much disk space a file with only one byte of content actually occupies. It begins by demonstrating that creating an empty file consumes no data blocks, only an inode. After writing a single character to the file, the directory's disk usage jumps to 4KB, revealing that even minimal data triggers allocation of a full block.

The explanation delves into the Linux ext2 filesystem structure, showing that the inode contains a block array (i_block) with 15 entries: 12 direct pointers, one single indirect, one double indirect, and one triple indirect. The block size on the author's system is 4KB, as confirmed by the dumpe2fs command. This means the smallest allocatable unit is 4KB, so a one-byte file still consumes an entire block.

The article further explains how large files are managed through indirect indexing, allowing exponential growth in addressable space. It concludes by noting that this block-based allocation principle applies to all filesystems, including distributed systems like HDFS, which uses 128MB blocks to reduce management overhead for large-scale data processing.

The key takeaway is that maintaining many small files is inefficient, as each consumes at least one full block plus inode overhead, regardless of actual content size.