Executing command inside container using docker-compose exec breaks mount from /sda2

[pinoniq]

I'm running the following docker-compose file:

version: '2'
services:
    web:
        image: pinoniq/httpd:1.4
        environment:
            - HOSTNAME=app.local
            - ENVIRONMENT=dev
            - FLAVOR=symfony
        ports:
            - "80:80"
        volumes:
            - ./api:/var/www

    php:
        image: pinoniq/php7:1.7
        ports:
            - "9000:9000"
            - "8080:8080"
        volumes:
            - ./api:/var/www
            - /var/www/node_modules
            - /var/www/vendor

    db:
        image: mariadb:10.1
        environment:
            - MYSQL_ROOT_PASSWORD=secret
        ports:
            - "3306:3306"

    rabbitmq:
        image: "rabbitmq:3-management"
        hostname: "rabbitmq"
        environment:
            RABBITMQ_ERLANG_COOKIE: "SWQOKODSQALRPCLNMEQG"
            RABBITMQ_DEFAULT_USER: "rabbitmq"
            RABBITMQ_DEFAULT_PASS: "rabbitmq"
            RABBITMQ_DEFAULT_VHOST: "/"
        ports:
            - "15672:15672"
            - "5672:5672"

both the http and php DockerFile can be fond here:
http: https://github.com/pinoniq/httpd
php: https://github.com/pinoniq/php7

The issue lies in the php docker container. More specifically the the /var/www/vendor mountpoint.

I do this to speed up development on windows. I mount the application under /var/www and then mount an additional internal mount to /var/www/vendor.

running docker-compose exec php mount | grep /var gives the following result:

//10.0.75.1/C on /var/www type cifs (rw,relatime,vers=3.02,sec=ntlmssp,cache=strict,username=jme,domain=EMAKINA,uid=0,noforceuid,gi
d=0,noforcegid,addr=10.0.75.1,file_mode=0755,dir_mode=0755,iocharset=utf8,nounix,serverino,mapposix,nobrl,mfsymlinks,noperm,rsize=1
048576,wsize=1048576,echo_interval=60,actimeo=1)
/dev/sda2 on /var/www/node_modules type ext4 (rw,relatime,data=ordered)
/dev/sda2 on /var/www/vendor type ext4 (rw,relatime,data=ordered)

All is fine and working nicely. I then however, get issues where after running a set of commands, the /var/www/vendor mount is removed.

The commands I'm running are:

# Update compose inside the docker container
docker-compose exec php composer update

# once finished, remove the local vendor file
rm -R -f ./api/vendor

# copy the docker vendor file to a mounted part of the container
docker-compose exec php cp -R /var/www/vendor /var/www/vendor-temp

# move the newly created local copy to vendor
mv ./api/vendor-temp ./api/vendor

It is most often the docker-compose exec php composer update command that makes the /var/www/vendor be removed.

Anyone any idea,

Host: Windows 10
docker version: Docker version 17.03.1-ce, build c6d412e
docker-compose version: docker-compose version 1.11.2, build f963d76

[shin-]

I think your volume configuration is an issue here. If /var/www/vendor gets removed/recreated at any point, it will remove the /var/www/vendor volume and replace it with a regular folder in the /var/www host bind, which seems like a reasonable behavior to me. I'd hit up the forums and explain what you're trying to achieve there where people can hopefully help you do it with a saner config.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[stale]

This issue has been automatically closed because it had not recent activity during the stale period.