OverlayFS Adds A "Volatile" Option - Faster Performance But All Syncs Are Omitted
OverlayFS as the Linux union mount file-system that allows combining multiple underlying mount points into one is seeing a new feature with Linux 5.10.
This decade-old alternative to the likes of AuFS and UnionFS is used by the likes of the OpenWrt project and also can be used for container/Docker use-cases too. The main new feature of the OverlayFS changes for Linux 5.10 is the introduction of a new "volatile" mode exposed via the new mount option of the same name.
The OverlayFS volatile mode is intended for select use-cases primarily around containers. The volatile mode aims to accelerate performance of the union file-system by omitting all sync calls to the upper file-system. Thus data security is at risk and as such the volatile mode is only intended where the data can be re-created in case of crash where data may be lost. But by avoiding syncs to the upper file-system, the performance can be much improved under heavy I/O operations.
This feature by means of touching a directory ($workdir/work/incompact/volatile) can indicate if a mount previously relied on the volatile mode so at the user's choosing could decide to toss away the data or otherwise continue with the mount and usage should it be known the data is safe.
The pull request of OverlayFS changes for Linux 5.10 also includes ioctl improvements, continued work in striving for unprivileged overlay mounts, and other enhancements.
This decade-old alternative to the likes of AuFS and UnionFS is used by the likes of the OpenWrt project and also can be used for container/Docker use-cases too. The main new feature of the OverlayFS changes for Linux 5.10 is the introduction of a new "volatile" mode exposed via the new mount option of the same name.
The OverlayFS volatile mode is intended for select use-cases primarily around containers. The volatile mode aims to accelerate performance of the union file-system by omitting all sync calls to the upper file-system. Thus data security is at risk and as such the volatile mode is only intended where the data can be re-created in case of crash where data may be lost. But by avoiding syncs to the upper file-system, the performance can be much improved under heavy I/O operations.
This feature by means of touching a directory ($workdir/work/incompact/volatile) can indicate if a mount previously relied on the volatile mode so at the user's choosing could decide to toss away the data or otherwise continue with the mount and usage should it be known the data is safe.
The pull request of OverlayFS changes for Linux 5.10 also includes ioctl improvements, continued work in striving for unprivileged overlay mounts, and other enhancements.
5 Comments