Is it possible to union-mount an existing disk image containing a root filesystem to the upperdir of an overlayfs mount for read-write access? I'm trying to union-mount some old disk images to a /tmp/ directory on an Ubuntu 16.04 machine using pre-existing scripts (originally written on Ubuntu 12.04 and ported to 14.04). The old automation scripts used overlayfs to create and union-mount writable ext2/3 disk images on top of squashfs images — something like this: cd /tmp mkdir lower upper union losetup /dev/loop1 /tmp/image.sfs losetup /dev/loop2 /tmp/image.ext2 mount -t squashfs -o ro /dev/loop1 /tmp/lower mount -t ext2 -o rw /dev/loop2 /tmp/upper mount -t overlayfs -o rw,lowerdir=/tmp/lower,upperdir=/tmp/upper overlayfs /tmp/union They were last used on an 3.x kernel (likely pre-3.18), and that method doesn't seem to work anymore. Overlay now requires a workdir option—e.g.: mount -t overlay -o rw,lowerdir=/tmp/lower,upperdir=/tmp/upper,workdir=/tmp/work overlay /tmp/union **Q:** Is it possible to tweak the automation scripts to mount and use the existing images in a writable state on a 4.x kernel? Can I avoid reformulating them to contain something like root and work directories? In some cases that may break their use elsewhere. Per the documentation, the workdir has to be an empty directory on the same filesystem as the upperdir to allow for atomic writes. That sounds to me like it is impossible to union-mount root filesystem images as read-write. The workdir would have to exist within the images (separate from their data directory) which isn't possible in a root filesystem image. I've considered loading up stock Ubuntu 14.04 in a VM for a couple one-offs, but it's not a long term solution. --- **Best Attempt So Far:** The only solution I've come up with that avoids double-writing the same data to disk is to mount both images as lowerdir values, with upperdir and workdir directories sitting on a tmpfs mount. Then I can use rsync to copy the changes from the tmpfs location to the mounted ext2/3 image after the automated writes are complete and the overlay union has been unmounted—i.e.: mount -t overlay -o rw,lowerdir=/tmp/upper:/tmp/lower,upperdir=/tmp/tmpfs/root,workdir=/tmp/tmpfs/work overlay /tmp/union ...perform automated reads/writes... umount /tmp/union ...rsync contents from /tmp/tmpfs/root to /tmp/upper... It's strictly limited to available RAM, though, and it's an ugly hack to add to all of the automation scripts.

I tried to mount an external storage device (a usb flash drive) with overlay (as the lowerdir the usb and as upper a temporary ramdisk) mkdir /data mkdir /data/lower mkdir /data/rw mkdir /data/root Then creating the ramdisk mount -t tmpfs none /data/rw mkdir /data/rw/upper mkdir /data/rw/work I mounted the usb device mount -t vfat /dev/sdb1 /data/lower To finally use overlay : mount -t overlay -o lowerdir=/data/lower,upperdir=/data/rw/upper,workdi=/data/rw/work none /data/root Which give me the folowwing error > wrong fs type, bad option, bad superblock on overlay, missing codepage or helper program, or other error > overlayfs: filesystem on '/data/lower' not supported But according to the documentation > The lower filesystem can be any filesystem supported by Linux and does not need to be writable Is there a link with automount and if so how I can fix it ?

Context: Say I have two workplaces. Currently, I carry a laptop along, and have external HDs staying in these workplaces to hold backups. Now, I'm given the opportunity to have a desktop in both workplaces, so I'd rather carry along an external HD (USB SSD) holding my files, and use the internal HDs of the desktops to hold backups. The desktops don't have the same hardware, so I'd keep config files specific, thus not store them on the external HD. The external HD would only hold ~/Documents ~/Images etc. The plan is to have /home mounted this way: unionfs /HomeLocal=RW:/mnt/external_drive/HomeRoaming=RW /home /HomeLocal will hold dotfiles and dotdirs. How should I instruct systemd to mount /home this way before users may log in? This is Linux Mint.

I would like to be able to merge changes from the upper filesystem of an overlay mount to a lower filesystem. I am interested both **online** (i.e. merge *while* the overlay is mounted) and **offline** (unmount the overlay and then merge) solutions. I have found a couple of offline solutions, which I have added as answers. Does anyone know of any **online** solutions? It would be good to have a "commit" type command you could run to merge down the layers while the overlay is still mounted. Something like this has been asked in the following questions with no answer: - Is there anyway to modify one file and sync to the lower directory in OverlayFs? - How to modify a lower file through an OverlayFS directory? - OverlayFS Seamlessly Edit File in Lower Directory - OverlayFS - How to make changes to upper filesystem persistent without unmounting? Comments in these posts suggest variously mergerfs and bcache , both of which solve specific use cases but not the generic filesytem-agnostic use case that overlays provide. My goal is to have a safe filesystem sandbox with snapshots that can be used with any Linux application over any (where practical) underlying filesystem, allowing you to roll back changes or manually commit them when you are ready. I have a suspicion that modern mainline Linux has all the necessary features to do this built-in, thanks to all the sandboxing/virtualization innovations of the last few years, if only I knew how to use them.

Instead of just mounting tmpfs on /var/log I want to use overlayfs. 1. /var/log are writable tmpfs, but containing files were there before tmpfs mount. This old files are not in memory of tmpfs but in lower layer. 2. only changes are stored in tmpfs, while old and unmodified files stored on SSD 3. sometimes it should be possible to write changes to SSD, for example via cron. This should free up tmpfs memory So, result should be: logs written to RAM, old and new boot logs accesable via same path. Changes are written sometimes to disk, by script. Point is to speed up a little, and safe SSD from many writes. (_I saw similar thing in puppy linux, not for logs, but for all changes to root_, but without installing it can't do the same, documentation not helps) I will do same for browser cookies/cache based on answer. But persistent write will be done on browser close. Can't turn off browser cache, need at least small cache to have same bugs in my web development as users can have because of cache.

AFAIK with [aufs](https://en.wikipedia.org/wiki/Aufs) when a file is touched (ie. only it's metadata changed, not it's contents), then the copy-on-write mechanism of the file system will copy the whole file into the r/w layer, essentially duplicating the file. Say, I only want to change the **metadata** (modification time, permissions, owner, etc.) of a *lot* of files, is there some union-type file system (OverlayFS perhaps?) that only keeps track of the metadata changes, unless the file *content* is altered - resulting in much less duplication? Or maybe some other trick can lead to the desired result?

I have searched the whole internet and have watched my lecture multiple times just to understand one thing: I still can't quite understand, how is overlayFS a solution to copying the central FS for each user? Is the lowerdir set to be the root directory? Or do we create an alternative rootfs and use it everywhere as a lowerdir for each new filesystem for each user? I apologize but there's some understanding gap. I don't understand in what way are we saving storage, and if someone could just put it in simple terms, it would be wonderful. Thank you!

I run openmediavault 4.1 and changed from RAID6 to SnapRaid with 7 hard drives for data plus parity. I remember reading installation advice on RAID setup that said "never set spindown for HDD" while using autoshutdown. So I was running fine for some years having the whole system shut down after 30 minutes of beeing idle - instead of having the disks beeing spun down and leaving the rest on power consumption. Now I changed my setup to SnapRaid with mergerfs and I wonder if I have to reset my spindown settings in Power Management. Access on a RAID array spins up all of the disks at once. With SnapRaid only the one disk with the actual data on it will be activated. So it would make sense to put all of them to spindown because most of them will not be used anyway. I decided so put it on "minimum consumption WITH spindown" and set spindown to 10 minutes. The whole system shuts down after 60 minutes of beeing ide. I think this is better because SnapRaid will spin up only the one disk it needs if there is some access. Is this a good setup?

After building Linux From Scratch several times I realized that package management is really a hard thing to do. After thinking for some time, I came across an idea. My idea is to install all packages into separate directories, then combine them into one file system and mount as root (required). Also, I want to attach and detach packages without a reboot (optional, preferred). How can I do this from scratch (or with a package)? (No problem if I need to program (C or C++ preferred))

The [overlayfs documentation](https://www.kernel.org/doc/html/latest/filesystems/overlayfs.html) states that "if the underlying filesystem is changed, the behavior of the overlay is undefined, though it will not result in a crash or deadlock." And, indeed, I've observed that when I make changes to the lowerdir in an overlay mount, strange stuff happens. However, if I remount afterwards, the problems seem to be fixed. If I ignore the documentation's advice and modify the underlying filesystem online, will mount -oremount merged *necessarily* fix any issues caused by doing so? (Yes, yes, I know the documentation says "changes to the underlying filesystems while part of a mounted overlay filesystem are **not allowed**" (emphasis mine), but I've heard [otherwise](https://unix.stackexchange.com/a/386758/173715).)

I have been using union mounts, such as unionfs on FreeBSD or mergerfs on Linux, and now I would need to do the same on illumos (SmartOS). Reading [illumos](https://illumos.org/man/) or [SmartOS](https://smartos.org/man/) man pages is not helpful. Same with searching them [via external search engine](https://www.google.com/search?hl=en&q=site%3Ahttps%3A%2F%2Fillumos.org%2Fman%2F) . If possible, how can one create a union mount (bind two or more directories to one mountpoint) on illumos / solaris? If not possible, is there anything close to same effect? ## Why In my humble opinion, union-like filesystems are used in relatively complex scenarios. Describing some here would be out of scope for this question. I would like to know if this is possible _a-priori_ in any case, as I would not want an answer solving only a subset of all unionfs possible use-cases. As an example: I need to pick and choose _from the subdirectories_ of read-only zfs datasets, mounted DVD and readonly USB drives, overlaid with read-write datasets. The resulting mount point needs to be available as a cifs share. It would need to be done programmatically after boot, depending on various system factors. For the interested, see referenced articles with some use cases. There are many such sysadmin scenarios. --- https://whattheserver.com/what-are-unionfs-mounts-and-how-to-use-them/ https://medium.com/@knoldus/unionfs-a-file-system-of-a-container-2136cd11a779

I have a folder containing certain files. Is there a way to mount it onto /, so that the contents of a given directory appear in the corresponding locations in /? In my case, the directories contain the files related to an application -- it does not have directories like /dev and /proc. I tried creating a bind mount in the following manner: mount --bind ~/applications/firefox / However, this does not do anything at all. Unlike what has been speculated in the comments, it does not have any adverse effect either. --- The motivation for this is that I have a number of custom applications that I want to install cleanly on mission-critical systems. Understandbly, I want to keep the related management to a minimum. I obtain the filesystem layout of the application install with a [unionfs-fuse + chroot method](https://unix.stackexchange.com/questions/140599/chroot-jail-that-can-access-the-filesystem) , but then I still have to deal with the problem of the applications' files mingling with those of the system. Mounting certain directories onto / would help me to keep the filesystem clean. While this method of installing software may seem perverted, there is a distro named "Porteus" which installs software in exactly this manner (and for the same reasons), however, they are known to use a patched kernel along with aufs. However, I cannot use a custom kernel on the systems.

A [union mount](https://en.wikipedia.org/wiki/Union_mount) presents a combined view of multiple directories (branches) in a single hierarchy. Ok, but how do I use one in practice, and which one? Many union mount technologies are available on a modern Linux system: [OverlayFS](https://en.wikipedia.org/wiki/OverlayFS) , [aufs](https://en.wikipedia.org/wiki/Union_mount) , [UnionFS](https://en.wikipedia.org/wiki/UnionFS) , various FUSE-based filesystems, … Given a use case, how do I decide which one(s) are suitable? In particular: * Which ones require root access on a typical system? (In other words, can an unprivileged user create a union mount?) * Which ones can be mounted on / (with root access of course)? * Which ones support writing to the underlying branches? (I.e. if /foo and /bar are union-mounted to /union, can I write to /foo and /bar?) * Which ones support configurable policies for writes? (E.g. can I sent newly created files to branch B even though I'm getting content from branch A when a file exists on both branches?) * Which ones support a priority order between branches that depends on the files? (E.g. can I always get the newest file among all the branches that have it?) Usage examples would be appreciated. Please avoid answers that focus on a specific point. I'm looking for comprehensive answers that review the available software (something similar to https://unix.stackexchange.com/questions/198590/what-is-a-bind-mount) . Good answers will be bountied.

From man mount_unionfs: If a file exists in the upper layer then there is no way to access a file with the same name in the lower layer. If necessary, a combination of loopback and union mounts can be made which will still allow the lower files to be accessed by a different pathname. How do I do this? I can't find a reference for file system loopback other than mounting .ISO and other block images which is clearly not the right meaning, and nullfs which just mentions the term in passing (but no concrete information). # My guess as to what's involved: I think it might involve first mounting the "lower" layer (possibly twice, to 2 different mountpoints?), then mounting a unionfs over that mountpoint (or over one of the two mountpoints) to a second mountpoint. The idea being that the same file system is exposed unmodified through the first mountpoint and also exposed as a lower layer of a union via the second mountpoint. This is based off a hint at this web-page . **Update, clarifying the aim in response to @arrowd:** The man page says one can access the lower files (implied: the entire set of lower files) via loopback, and that this can makes the lower layer accessible in parallel both (A) as a lower layer of the unionfs, and (B) via a different "loopback" mount-point/pathname as a file system in its own right. I understand the concept of a loopback layer so this makes sense, and the man page says it's doable, but I can't find the commands that will let me do it. Being more specific, I want to overlay a file system in the usual way as a unionfs, but I also want a separate mountpoint to expose the lower layer by itself, to allow some tasks to read/write/search the entire "lower" file system *directly* as well, *without* being obscured/masked/overlaid by the union "overlay fs". So it isn't just about a single file - I need the entire "lower" layer also accessible via loopback and a separate mountpoint, not just accessible through the unionfs layer or on a file-by-file basis. It probably means I need to mount it twice in some way or another, first via loopback and then via unionfs, at the point where I initially set up/mount the unionfs layer, but I don't know how to do it. The aim is that some cron tasks (and occasionally I as well) will need to update the files within the lower layer (without unmounting/disturbing the unionfs overlay), and **not** not just copy and update them within the "upper" layer as would happen with simple use of unionfs. So I need to have two mountpoints in use - one exposing the lower layer overlaid with a unionfs, and the other exposing the lower layer only **without** the unionfs overlay, and be able to use both in parallel in different tasks. Also see comment I added above - I *think* it might involve mounting the lower file system twice (once directly, once as union), or mounting it once and remounting from that mountpoint a second time (as union). But I'm not sure and also unclear how safe it is. Hopefully this is more clear as to the aim of the question.

From man mount_unionfs : BUGS: THIS FILE SYSTEM TYPE IS NOT YET FULLY SUPPORTED (READ: IT DOESN'T WORK) AND USING IT MAY, IN FACT, DESTROY DATA ON YOUR SYSTEM. USE AT YOUR OWN RISK. BEWARE OF DOG. SLIPPERY WHEN WET. BATTERIES NOT INCLUDED. This code also needs an owner in order to be less dangerous Bit scarey eh? It goes on to explain some of the issues but it would need specialist knowledge of the file system handling to be sure what the extent of issues and edge cases are. Nonetheless perhaps it is well known, or clear to someone with more knowledge: Are the dangers well enough defined that it's only under certain uses, or certain less-likely/easily avoided edge-cases, that a serious issue might arise? Or is closer to being *"if your data matters never, ever, ever, use unionfs mount with it, even with a gun to your head"*

What's an easy way to to use a union filesystem (read-only SquashFS + R/W overlay) as /? I'm thinking of using a custom boot script as the kernel init= argument. The script would mount the SquashFS + overlay, then chroot into the mounted filesystem and run the real /sbin/init... however I'm not sure how well this works with systemd init, and I'm wondering if there isn't a simpler option.

Is it possible to do an overlay mount when one of the paths has a colon in it? All of the FUSE overlay mounting solutions I've looked at use a colon to separate the paths in the overlay, and I can't find a way to escape it.

Say, we initially have / with the following contents: /bin/foo /usr/bin/bar /usr/lib/libfoo.so We also have another directory, say /stuff, with the contents: usr/lib/mystuff/mystuff.so etc/foo/qux/buzz bin/stuff Is there a way to create a union between / and /stuff, so that I get: /bin/foo /bin/stuff /usr/bin/bar /usr/lib/libfoo.so /usr/lib/mystuff/mystuff.so /etc/foo/qux/buzz (with all of the content from /stuff there as well, just not shown there.) This union should be of a **temporary nature** (just like filesystem mounts are), otherwise I could have used cp, mv or rsync as well. How should I go about doing this? _One would expect that an answer to this would be easy to find; and indeed, there are many online resources that mention creating unions between two directories and mount them elsewhere, but none really cover this particular case of mounting a directory "over" /._

Here's the test case: #!/bin/bash more="$@" saved="$EUID" if [ "$EUID" = "0" ]; then echo "don't run as root" exit fi m="/tmp/mount-point" exec sudo unshare -m /bin/bash </proc, as you can probably guess, is mounted and not empty in the _big fs_.

I've grown up a Debian/Ubuntu guy and I'm learning Fedora by trying to recreate the services my Ubuntu trusty box currently provides. However, I'm trying to recreate a union mount and don't see any obvious way to do it. Each user on my Ubuntu-based server has a "media" folder, with subfolders for movies, music, etc. Using aufs-tools, I create a (read-only) union mount of userA...userZ's "media" folders at /srv/media. Then using Samba, my Plex box can access that combined media folder over the network. It appears nothing reasonably resembling aufs-tools is available via the standard repos (nor via rpmfusion). I've seen some mentions online of UnionFS, but I also saw something about it having fallen out of favor. I don't want to use FUSE, because this union mount is based on several users' media folders. What's the preferred solution for creating union mounts on Fedora/RHEL/CentOS servers these days?