Populate user dir once (getgrav#44)

I found another way to install grav by populating user dir if empty during container start.
Therefore i skip user dir during build. A script will then populate /var/www/user as described in README.md
This commit also includes https://github.com/Freiheitswolke/docker-grav/tree/fix/volume and some changes to README.md to reflect the changed volumes.
The commit will not break the old way with a mount for the whole '/var/www/html' for those who want to upgrade grav via the admin console.
Building or pulling the container image with the new version of grav is a more proper way to perform upgrades in my opinion.

Closes getgrav#44

Dockerfile

@@ -54,11 +54,11 @@ USER www-data
 # Define Grav specific version of Grav or use latest stable
 ARG GRAV_VERSION=latest
 
-# Install grav
+# Install grav (skip user dir)
 WORKDIR /var/www
 RUN curl -o grav-admin.zip -SL https://getgrav.org/download/core/grav-admin/${GRAV_VERSION} && \
     unzip grav-admin.zip && \
-    mv -T /var/www/grav-admin /var/www/html && \
+    for file in $(ls -A "grav-admin/" | egrep -v user); do mv "grav-admin/$file" "html/"; done && \
     rm grav-admin.zip
 
 # Create cron job for Grav maintenance scripts
@@ -70,9 +70,21 @@ USER root
 # Copy init scripts
 # COPY docker-entrypoint.sh /entrypoint.sh
 
+# Create script that populates user dir
+RUN { \
+    echo '#!/bin/bash'; \
+    echo '# Ensure user dir exists'; \
+    echo 'sudo -iu www-data mkdir -p /var/www/html/user/'; \
+    echo '# Populate user dir if it is empty'; \
+    echo 'if test -n "$(find /var/www/html/user -maxdepth 0 -empty)" ; then'; \
+    echo 'cp -rp /var/www/grav-admin/user/* /var/www/html/user/'; \
+    echo 'fi'; \
+    } > /usr/local/bin/populate-userdir && \
+    chmod +x /usr/local/bin/populate-userdir
+
 # provide container inside image for data persistence
-VOLUME ["/var/www/html"]
+VOLUME ["/var/www/html/backup", "/var/www/html/logs", "/var/www/html/user"]
 
 # ENTRYPOINT ["/entrypoint.sh"]
 # CMD ["apache2-foreground"]
-CMD ["sh", "-c", "cron && apache2-foreground"]
+CMD ["sh", "-c", "populate-userdir && cron && apache2-foreground"]

README.md

@@ -14,9 +14,11 @@ This currently is pretty minimal and uses:
 
 ## Persisting data
 
-To save the Grav site data to the host file system (so that it persists even after the container has been removed), simply map the container's `/var/www/html` directory to a named Docker volume or to a directory on the host.
+To save the Grav site data to the host file system (so that it persists even after the container has been removed), simply map the container's `/var/www/html/user` directory to a named Docker volume or to a directory on the host.
 
-> If the mapped directory or named volume is empty, it will be automatically populated with a fresh install of Grav the first time that the container starts. However, once the directory/volume has been populated, the data will persist and will not be overwritten the next time the container starts.
+> If the mapped directory or named volume is empty, it will be automatically populated with a fresh user dir the first time that the container starts. However, once the directory/volume has been populated, the data will persist and will not be overwritten the next time the container starts.
+
+This also applies for '/var/www/backup' and '/var/www/logs' except population at first startup.
 
 ## Building the image from Dockerfile
 
@@ -35,7 +37,7 @@ Point browser to `http://localhost:8000` and create user account...
 ## Running Grav Image with Latest Grav + Admin with a named volume (can be used in production)
 
 ```
-docker run -d -p 8000:80 --restart always -v grav_data:/var/www/html grav:latest
+docker run -d -p 8000:80 --restart always -v grav_backup:/var/www/backup -v grav_logs:/var/www/logs -v gravi_user:/var/www/html/user grav:latest
 ```
 
 ## Running Grav Image with docker-compose and a volume mapped to a local directory