A bind mountbind mountA host file or directory mounted at a path inside a container makes a host path appear inside a container. Both sides see the same files: no copy, no sync.

  host                                    container "docs"
  /home/learner/site/index.html  ◄────►  /usr/share/nginx/html/index.html
                                 same file

Volume or bind mount

-v treats the part before the colon as a host path when it starts with /, ~ or ., and as a volume name otherwise.

Serve a project directory

docker run -d --name docs -p 8080:80 -v ~/site:/usr/share/nginx/html nginx:1.27
curl -s localhost:8080

<h1>Release notes</h1>

Edit on the host and request again. No restart:

echo '<p>v2.1: faster uploads</p>' >> ~/site/index.html
curl -s localhost:8080

<h1>Release notes</h1>
<p>v2.1: faster uploads</p>

This is the usual development loop: the editor writes to the host, the container serves or runs the result.

Files the container writes are root-owned

Processes in most containers run as root (uid 0). Files they create on a bind mount keep uid 0 on the host:

docker exec docs sh -c 'date > /usr/share/nginx/html/build.txt'
ls -l ~/site

total 8
-rw-r--r-- 1 root    root    29 Oct 11 20:44 build.txt
-rw-r--r-- 1 learner learner 51 Oct 11 20:44 index.html

learner can delete build.txt (deleting depends on the directory, which learner owns) but cannot edit it:

$ echo x >> ~/site/build.txt
bash: /home/learner/site/build.txt: Permission denied

Take ownership back with sudo:

sudo chown learner:learner ~/site/build.txt

On real projects the longer-term fix is to run the container as your own uid with docker run --user, so its files are yours from the start.