Allowing multiple services in docker-compose to share a merged volume - docker

Given a docker-compose.yml file like below, I'm looking for a way that both service a and b can have access to a shared volume which consists of the merged contents of both containers.
version: '3'
volumes:
shared-merged-volume:
services:
a:
volumes:
- shared-merged-volume:/shared
b:
volumes:
- shared-merged-volume:/shared
Let's say service a has a directory at /shared/dir-from-a and service b has a similar /shared-dir-from-b directory. The desired result is to end up with:
$ ls /shared # from either container
dir-from-a
dir-from-b
What I find is that one of the containers "wins" and only one of those two directories is ever present. I can work around the issue like this but is more verbose and requires modification if directory contents ever changes:
version: '3'
volumes:
service-a-shared-volume:
service-b-shared-volume:
services:
a:
volumes:
- service-a-shared-volume:/shared/dir-from-a
- service-b-shared-volume:/shared/dir-from-b
b:
volumes:
- service-a-shared-volume:/shared/dir-from-a
- service-b-shared-volume:/shared/dir-from-b
Thanks in advance for any help!

Is using a named volume a requirement?
If not, then to accomplish such merging I usually just map directories to one location on the host drive, instead of using volumes, and it merges with no problems. Tested on big loads and multiple containers writing simultaneously.
proposed compose file:
version: '3'
volumes:
shared-merged-volume:
services:
a:
volumes:
- /location/on/host/system:/shared
b:
volumes:
- /location/on/host/system:/shared
Edit from comments
This method mounts everything that's in the local host directory to the /shared, meaning if it's empty - it'll mount the empty dir, and whatever was there - will be overwritten by the empty dir. Everything that will be written inside that mount after your service starts, will be persisted and merged across services as expected.

If both containers are creating different folders, I don't see how they can be contending to create their own respective folders, unless they both delete the contents of /shared first, then they create the folders? But that would mean that the use of volumes in this case is null because the contents will be deleted every time the container starts?
In any case, I find that it is often useful to persuade the containers to share the same folder by use of path redirection. I will share two ways of accomplishing this:
If you have access to the code that creates the folders in /shared, then you can use environment variables to change the expected location of /shared for each service
version: '3'
volumes:
shared-merged-volume:
services:
a:
environment:
SHARED_VOLUME_PATH: /shared/a/
volumes:
- shared-merged-volume:/shared
b:
environment:
SHARED_VOLUME_PATH: /shared/b/
volumes:
- shared-merged-volume:/shared
You may need to have the services create SHARED_VOLUME_PATH, but now they can both live peaceably with each other.
If you are unable to change the location of /shared, which means each service will always want to use that path, another way to create path redirection is to use symbolic links. For this to work, you will have to override the entrypoint of your services or do this step during the build process of the image.
version: '3'
volumes:
shared-merged-volume:
services:
a:
entrypoint: [ "ln", "-sf", "/symshared/a/", "/shared/" ]
volumes:
- shared-merged-volume:/symshared
b:
entrypoint: [ "ln", "-sf", "/symshared/b/", "/shared/" ]
volumes:
- shared-merged-volume:/symshared
Alternatively, build the images ahead of time, and add a simple RUN command in the Dockerfile which creates this symbolic link:
...
ARG SHARED_VOLUME_PATH
RUN ln -sf ${SHARED_VOLUME_PATH} /shared/
What this allows you to do is that each container will keep using /shared as they used to, but you will still be able to store it's content in the volume, without interfering with what other containers want to do with their own version of /shared.
Needless to say, the ln command only works on linux and other unixes, and in some cases, you may need to install it prior. If your container image is based on something else like windows for example, then find something else that can be used to create symlinks.

Related

Map the docker compose volume from container to host is not working

I have a very simple nextjs application where I have two folders which I like to map to the host (developer system) while I deploy this application inside docker (I use docker-desktop).
Data folder (It has some json files and also some nested folders and files)
Public folder (It has nested folders too but It contains images)
I have tested in local and also inside the docker container (without volume and all) - It's all functioning.
As a next step - I want to use the Volume with my docker-compose file so that, I can bind these directories inside the container with the source (and going forward with AKS file storage options).
I have tried with multiple approaches (also checked some of the answers in stackoverflow) but it does not help to achieve the same result.
Here is my docker-compose file for your reference.
version: '3.4'
services:
portfolio:
image: ${DOCKER_REGISTRY-}brij1111-portfolio
build:
context: ./APP-03/clientapp
dockerfile: dockerfile
volumes:
- /app/node_modules
# anonymous volume only for node_modules
- portfolio_data:/app/data
# named volume inside which the nextjs app writes content to the file
volumes:
portfolio_data:
driver: local
driver_opts:
o: bind
type: none
device: /APP-03/clientapp/data
# I have tried here to give a full path like /mnt/c/work/.../APP-03/clientapp/data but that also not working.
using docker-desktop I can see the volume indeed created for the container and it has all the files. However, It does not get reflected in my source if anything is updated inside that volume (like I add some content through nextjs application to that file it does not get reflected inside the running container).
in case, someone wants to know my folder hierarchy and where am i running docker-compose file, here is that reference image.
I had a similar problem installing Gitea on my NAS until someone (thankfully) told me a way to compromise (ie. your data will be persistent, but not in the location of your choosing).
version: '3.4'
volumes:
portfolio_data: {}
services:
portfolio:
image: ${DOCKER_REGISTRY-}brij1111-portfolio
build:
context: ./APP-03/clientapp
dockerfile: dockerfile
volumes:
- /app/node_modules
# anonymous volume only for node_modules
- portfolio_data:/app/data
In my articular case, I had to access my NAS using terminal, to the location which the container image is located, and search from there.
Hope it helps you

define volumes in docker-compose.yaml

I am writing a docker-compose.yaml file for my project. I have checked the volumes documentation here .
I also understand the concept of volume in docker that I can mount a volume e.g. -v my-data/:/var/lib/db where my-data/ is a directory on my host machine while /var/lib/db is the path inside database container.
My confuse is with the link I put above. There it has the following sample:
version: "3.9"
services:
db:
image: db
volumes:
- data-volume:/var/lib/db
backup:
image: backup-service
volumes:
- data-volume:/var/lib/backup/data
volumes:
data-volume:
I wonder does it mean that I have to create a directory named data-volume on my host machine? What if I have a directory on my machine with path temp/my-data/ and I want to mount that path to the database container /var/lib/db ? Should I do something like below?
version: "3.9"
services:
db:
image: db
volumes:
- temp/my-data/:/var/lib/db
volumes:
temp/my-data/:
My main confusion is the volumes: section at the bottom, I am not sure whether the volume name should be the path of my directory or should be just literally a name I give & if it is the latter case then how could the given name be mapped with temp/my-data/ on my machine? The sample doesn't indicate that & is ambiguous to clarify that.
Could someone please clarify it for me?
P.S. I tried with above docker-compose I guessed, ended up with the error:
ERROR: The Compose file './docker-compose.yaml' is invalid because:
volumes value 'temp/my-data/' does not match any of the regexes: '^[a-zA-Z0-9._-]+$'
Mapped volumes can either be files/directories on the host machine (sometimes called bind mounts in the documentation) or they can be docker volumes that can be managed using docker volume commands.
The volumes: section in a docker-compose file specify docker volumes, i.e. not files/directories. The first docker-compose in your post uses such a volume.
If you want to map a file or directory (like in your last docker-compose file), you don't need to specify anything in the volumes: section.
Docker volumes (the ones specified in the volumes: section or created using docker volume create) are of course also stored somewhere on your host computer, but docker manages that and you shouldn't normally need to know where or what the format is.
This part of the documentation is pretty good about explaining it, I think https://docs.docker.com/storage/volumes/
As #HansKilian mentions, you don't need both volumes and services.volumes. To use services.volumes, map the host directory to the container directory like this:
services:
db:
image: db
volumes:
- /host/path/lib/db:/container/path/lib/db
With that, the directory /host/path/lib/db on the host machine will be used by the container and available at /container/path/lib/db.
Now, if you're like me, I get really confused with fake examples, so let's say the real directory on your host machine is /var/lib/db and you just want to see it at /db when you run a shell in Docker (i.e., docker exec -it /bin/bash container-id).
docker-compose.yaml would look like this:
services:
db:
image: db
volumes:
- /var/lib/db:/db
Now when you run the shell, cd /logs and ls, you'll see the same results as if you'd cd /var/lib/db on the host.
If you want to use the volumes section to indicate a global volume to use, you first have to create that volume using docker volume create. The documentation Hans linked includes steps to do this. The syntax of /host/path:/container/path is replaced by volume-name:/container/path. Then, once defined, you'd alter your docker-compose.yaml to be more like this:
services:
db:
image: db
volumes:
- your-global-volume-name:/db
volumes:
your-global-volume-name:
external: true
Note that I have not tested or used the this configuration. I'm assuming it's correct based on the other method working and the few changes I can identify in the docs.

Docker change location of named volumes

I have a problem that I just can't understand. I am using docker to run certain containers, but I have problems with at least one Volume, where I't like to ask if anybody can give me a hint what I am doing wrong. I am using Nifi-Ingestion as example, but it affects even more container volumes.
First, let's talk about the versions I use:
Docker version 19.03.8, build afacb8b7f0
docker-compose version 1.27.4, build 40524192
Ubuntu 20.04.1 LTS
Now, let's show the volume in my working docker-compose-file:
In my container, it is configured as followed:
volumes:
- nifi-ingestion-conf:/opt/nifi/nifi-current/conf
Below my docker-compose file it is defined as a normal named volume:
volumes:
nifi-ingestion-conf:
This is a snippet from the docker-compose that I'd like to get working
In my container, it is configured in this case as followed (having my STORAGE_VOLUME_PATH defined as /mnt/storage/docker_data):
volumes:
- ${STORAGE_VOLUME_PATH}/nifi-ingestion-conf:/opt/nifi/nifi-current/conf
On the bottom I guess there is something to do but I don't know what I could need to do here. In this case it is the same as in the working docker-compose:
volumes:
nifi-ingestion-conf:
So, now whats my problem?
I have two docker-compose files. One uses the normal named volumes, and one uses the volumes in my extra mount path. When I run the containers, the volumes seem to work different since files are written in the first style, but not in the second. My mount paths are generated in the second version so there is nothing wrong with my environment variables in the .env-file.
Hint: the /mnt/storage/docker_data is an NFS-mount but my machine has the full privileges on that share.
Here is my fstab-entry to mount that volume (maybe I have to set other options):
10.1.0.2:/docker/data /mnt/storage/docker_data nfs auto,rw
Bigger snippets
Here is a bigger snipped if the docker-compose (i need to cut and remove confident data, my problem is not that it does not work, it is only that the volume acts different. Everything for this one volume is in the code.):
version: "3"
services:
nifi-ingestion:
image: my image on my personal repo
container_name: nifi-ingestion
ports:
- 0000
labels:
- app-specivic
volumes:
- ${STORAGE_VOLUME_PATH}/nifi-ingestion-conf:/opt/nifi/nifi-current/conf
#working: - nifi-ingestion-conf:/opt/nifi/nifi-current/conf
environment:
- app-specivic
networks:
- cnetwork
volumes:
nifi-ingestion-conf:
networks:
cnetwork:
external: false
ipam:
driver: default
config:
- subnet: 192.168.1.0/24
And here of the env (only the value we are using)
STORAGE_VOLUME_PATH=/mnt/storage/docker_data
if i understand your question correctly, you wonder why the following docker-compose snippet works for you
version: "3"
services:
nifi-ingestion:
volumes:
- nifi-ingestion-conf:/opt/nifi/nifi-current/conf
volumes:
nifi-ingestion-conf:
and the following docker-compose snippet does not work for you
version: "3"
services:
nifi-ingestion:
volumes:
- ${STORAGE_VOLUME_PATH}/nifi-ingestion-conf:/opt/nifi/nifi-current/conf
what makes them different is how you use volumes. you need to differentiate between mount host paths and mount named volumes
You can mount a host path as part of a definition for a single service, and there is no need to define it in the top level volumes key.
But, if you want to reuse a volume across multiple services, then define a named volume in the top-level volumes key.
named volumes are managed by docker
If you start a container with a volume that does not yet exist, Docker creates the volume for you.
also, would advise you to read this answer
update:
you might also want to read about docker nfs volumes

How to create top-level named volume with a relative path for the source?

In docker-compose version 2, I set up one container with relative path mappings like:
volumes:
- ./local:/var/www/local
- ./another:/var/www/another
Then I used volumes_from for other services to reuse the same volume and paths.
Since compose version 3 removed volumes_from, I'm trying to create a named top-level volume that can be used for every service, but I don't see how to use a relative path. Since this is for local development, the absolute path can be different among developers.
The other option is to copy the above volumes configuration into every service that needs it, but as I understand it'll create a new volume for each one and add overhead for syncing. What's the solution here?
You can use driver options. Example
services:
web:
image: ubuntu
command: tail -f /dev/null
volumes:
- mydata:/test
volumes:
mydata:
driver: local
driver_opts:
o: bind
device: $PWD

File in docker-entrypoint-initdb.d never get executed when using docker compose

I'm using Docker Toolbox on Windows 10
I can access the php part succesfully via http://192.168.99.100:8000, I have been working around with the mariadb part but still having several problems
I have an sql file as /mariadb/initdb/abc.sql so I should be copied into /docker-entrypoint-initdb.d, after the container is created I use docker-compose exec mariadb to access the container, there is the file as /docker-entrypoint-initdb.d/abc.sql but the file never get executed, I also have tested to import the sql file to the container manually, it was succesful so the sql file is valid
I don't quite understand about the data folder mapping, and what to do to get the folder sync with the container, I always get the warning when recreate the container using docker-compose up -d
WARNING: Service "mariadb" is using volume "/var/lib/mysql" from the previous container. Host mapping "/.../mariadb/data" has no effect. Remove the existing containers (with docker-compose rm mariadb) to use the Recreating db ... done
Questions
How to get the sql file in /docker-entrypoint-initdb.d to be executed ?
What is the right way to map the data folder with the mariadb container ?
Please guide
Thanks
This is my docker-compose.yml
version: "3.2"
services:
php:
image: php:7.1-apache
container_name: web
restart: always
volumes:
- /.../php:/var/www/html
ports:
- "8000:80"
mariadb:
image: mariadb:latest
container_name: db
restart: always
environment:
- MYSQL_ROOT_PASSWORD=12345
volumes:
- /.../mariadb/initdb:/docker-entrypoint-initdb.d
- /.../mariadb/data:/var/lib/mysql
ports:
- "3306:3306"
For me the issue was the fact that Docker didn't clean up my mounted volumes from previous runs.
Doing a:
docker volume ls
Will list any volumes, and if previous exist, then run 'rm' command on the volume to remove it.
As stated on docker mysql docks, scripts in the '/docker-entrypoint-initdb.d' folder is only evalutated the first time the container runs, and if a previous volume remains, it won't run the scripts.
As for the mapping, you simply need to mount your script folder to the '/docker-entrypoint-initdb.d' folder in the image:
volumes:
- ./db/:/docker-entrypoint-initdb.d
I have a single script file in a folder named db, relative to my docker-compose file.
In your Docker file for creating mariaDB, at the end add the abc.sql file to your docker entry point like so:
COPY abc.sql /docker-entrypoint-initdb.d/
Remove the - /.../mariadb/initdb:/docker-entrypoint-initdb.d mapping as any file copied into the entry point will be executed.
Note: Windows containers do not execute anything in docker-entrypoint-initdb.d/

Resources