The other day I was trying to switch Docker’s storage driver from devicemapper to OverlayFS on my Debian Stretch development machine.
Docker’s instructions suggested first checking whether the machine has the
overlay kernel module:
$ lsmod | grep overlay
On my machine, this came up empty. Am I actually missing OverlayFS?! I’m on the latest vanilla Debian Stretch, with kernel version 4.6. OverlayFS was added to the kernel at version 3.18, released over two years ago.
In fact, Linux provides a command specifically to investigate your kernel modules:
$ /sbin/modinfo overlay filename: /lib/modules/4.6.0-1-amd64/kernel/fs/overlayfs/overlay.ko alias: fs-overlay license: GPL description: Overlay filesystem author: Miklos Szeredi <firstname.lastname@example.org> depends: intree: Y vermagic: 4.6.0-1-amd64 SMP mod_unload modversions parm: check_copy_up:bool parm: ovl_check_copy_up:Warn on copy-up when causing process also has a R/O fd open
Rad! So I’ve got OverlayFS after all. I can check that the kernel module loads successfully:
$ /sbin/modprobe overlay
No errors. We’re all good, and
lsmod confirms the module is loaded:
$ lsmod | grep overlay overlay 45056 11
In fact, the kernel waits to load the
overlay module until it’s needed, at which point it loads it automatically. So simply following the instructions from the Docker docs works without additional steps. The
lsmod | grep overlay check was just a red herring that caused me a false alarm.