Volume Modes: Filesystem vs Block
What you'll learn: How Kubernetes volumes can be presented to pods in two different modes—filesystem or block—and when to use each.
The Two Ways to Present Storage
When you attach storage to a pod, Kubernetes gives you two options for how that storage appears inside the container: as a filesystem (the default) or as a raw block device.
Filesystem Mode (volumeMode: Filesystem)
This is what you're probably used to. The storage volume gets formatted with a filesystem (like ext4 or xfs) and mounted at a specific path inside your container. Your application reads and writes files normally using standard file operations.
Think of it like: A USB drive that's already formatted—you plug it in, and your computer shows you folders and files you can work with immediately.
Block Mode (volumeMode: Block)
Here, the storage appears as a raw block device (like /dev/sda) without any filesystem. Your application talks directly to the raw storage device using special I/O operations. This is for specialized applications that need complete control over how data is written to disk.
Think of it like: An unformatted hard drive—advanced users can write data directly to specific sectors without the overhead of a filesystem layer.
When to Use Each
Use Filesystem mode (default) for:
- Most applications that work with files and directories
- Databases that manage files (like SQLite)
- Web servers serving static content
Use Block mode for:
- High-performance databases that implement their own storage management (like certain enterprise databases)
- Applications requiring direct disk I/O control
- Scenarios where filesystem overhead would hurt performance
Key Takeaway: Filesystem mode presents storage as a mounted directory (most common), while block mode exposes raw storage devices for applications needing direct, low-level disk access—choose based on your application's storage interface requirements.