Do I need to provide the docker instance with a real IP address instead of a loopback address 127.0.0.1?
Usually, it isn't necessary, and you should be able to connect with the default parameters which are localhost and 7687.
But if something is off, you can find the address of the Docker container with:
docker inspect -f '{{range.NetworkSettings.Networks}}{{.IPAddress}}{{end}}' container_name_or_id
And you can find the container_name_or_id with docker ps.
The connection command would look like this:
memgraph = Memgraph(host="ip_address", port=7687)
Related
I'm new with docker and I'm probably missing a lot, although i went through the basic documentation and I'm trying to deploy a simple Spring Boot API
I've deployed my API as a docker-spring-boot .jar file , then i installed docker and pushed it with the following commands:
sudo docker login
sudo docker tag docker-spring-boot phillalexakis/myfirstapi:01
sudo docker push phillalexakis/myfirstapi:01
Then i started the API with the docker run command:
sudo docker run -p 7777:8085 phillalexakis/myfirstapi:01
When i visit localhost:7777/hello I'm getting the desired response
This is my Dockerfile
FROM openjdk:8
ADD target/docker-spring-boot.jar docker-spring-boot.jar
EXPOSE 8085
ENTRYPOINT ["java","-jar","docker-spring-boot.jar"]
Based on this answered post this the command to get the ip address
docker inspect -f '{{range .NetworkSettings.Networks}}{{.IPAddress}}{{end}}' container_name_or_id
So, i run it with container_name_or_id = phillalexakis/myfirstapi:01 and I'm getting this error
Template parsing error: template: :1:24: executing "" at <.NetworkSettings.Networks>: map has no entry for key "NetworkSettings"
If i manage somehow to get the IP will i be able to visit it and get the same response?
This is how i have it in my mind: ip:7777/hello
You have used the image name and not the container name.
Get the container name by executing docker ps.
The container ID is the value in the first column, the container name is the value in the last column. You can use both.
Then, when you have the IP, you will be able to access your API at IP:8085/hello, not IP:7777/hello
The port 7777 is available on the Docker Host and maps to the port 8085 on the container. If you are accessing the container directly - which you do, when you use its IP address - you need to use the port that the container exposes.
There is also another alternative:
You can give the container a name when you start it by specifying the --name parameter:
sudo docker run -p 7777:8085 --name spring_api phillalexakis/myfirstapi:01
Now, from your Docker host, you can access your API by using that name: spring_api:8085/hello
You should never need to look up that IP address, and it often isn't useful.
If you're trying to call the service from outside Docker space, you've done the right thing: use the docker run -p option to publish its port to the host, and use the name of the host to access it. If you're trying to call it from another container, create a network, make sure to run both containers with a --net option pointing at that network, and they can reach other using the other's --name as a hostname, and the container-internal port the other service is listening on (-p options have no effect and aren't required).
The Docker-internal IP address just doesn't work in a variety of common situations. If you're on a different host, it will be unreachable. If your local Docker setup uses a virtual machine (Docker Machine, Docker for Mac, minikube, ...) you can't reach the IP address directly from the host. Even if it does work, when you delete and recreate the container, it's likely to change. Looking it up as you note also requires an additional (privileged) operation, which the docker run -p path avoids.
The invocation you have matches the docker inspect documentation (as #DanielHilgarth notes, make sure to run it on the container and not the image). In the specific situation where it will work (you are on the same native-Linux host as the container) you will need to use the unmapped port, e.g. http://172.17.0.2:8085/hello.
I want to start an Elasticsearch container in Docker. By default I see nearly everywhere something like:
docker run -d -p 9200:9200 docker.elastic.co/elasticsearch/elasticsearch:5.5.1
Now is my question: Why are we mapping the port on our host network? I understand port mapping but I don't see the big advantage of it.
In my opinion I would always do something like this:
$ docker network create logging
20aa4c7bf2d8289d8cbd485c3e384f9371eed87204625998687c61e4bad27f14
$ docker run -d --name es --net logging docker.elastic.co/elasticsearch/elasticsearch:5.5.1
And connect to the ES by using it's name (es in this case) and deploying containers in the same network. I would think my ES is more secure in its private docker network.
I see there is an advantage for port mapping when your containers which need to connect to elasticsearch aren't in the same network. But are there other advantages or why is this always shown with port mapping?
So host access is more about accessibility. If you are running docker on local machine and you want to access the app only on that machine, then host mapping is not need.
Now if you need to access this app on a external computer other than your docker host then you need to do that port mapping.
docker run -d -p 9200:9200 docker.elastic.co/elasticsearch/elasticsearch:5.5.1
This maps the host port 9200 (left side) to 9200 inside the docker (right side). The listening interface is 0.0.0.0 which means all interfaces. And hence it is accessible to to anyone how has access to this machine.
If you want to make it more secure then you do it like below
docker run -d -p 127.0.0.1:9200:9200 docker.elastic.co/elasticsearch/elasticsearch:5.5.1
This would listen on local host only. So only you can access it on the machine. But if you need to access it from some place else then you would use a SSH tunnel
ssh -L 9200:127.0.0.1:9200 <user>#<HOSTIP>
And on that machine you can access it on 127.0.0.1:9200
Next level of security is added when you use a firewall like ufw, firewalld etc.
What you did with network command
docker network create logging
Basically creates new network and isolates other docker containers from accessing it on the host. But as long as external accessibility is concerned, you still need to map it to the host port
Hope this answers your question
What could be wrong if the docker port cannot be accessed via the Access URL:
https://hub.docker.com/r/library/crate/
Since crate is listening on 127.0.0.1 it won't be accessible from outside the world since for the host 127.0.0.1 means something else. So you need to make your crate container listen on host IP. You can do that using:
$ docker run --net=host -d crate crate -Cnetwork.host=_site_
Details here. Also, there is change in release 2.X.X which is causing trouble for a lot of people it says:
Changed default bind and publish address from 0.0.0.0 to the system
loopback addresses which will result in CrateDB listening only to
local ports.
Have a look at release notes here.
I'm running a CentOS-based container on Docker for Windows and trying to connect to an http service running on port 8545 of my host environment.
I've tried this, attempting a variety of suspected host names and IP addresses:
curl http://localhost:8545
But the error message I get is "curl: (7) Failed connect to localhost:8545; Connection refused"
How should I figure out what IP Address to use? Is there anything I need to configure as far as allowing the port number to be accessed from inside the container?
Localhost is not working yet I think with Docker for Windows.
There is few things you can try. First you can add EXPOSE 'portnumber' in the dockerfile so the container will listen on this port. You can also use docker run with -p 8545:8545, it will map the port of the container and the host.
To get the Ip address of the container you can use:
docker inspect -f "{{ .NetworkSettings.Networks.nat.IPAddress }}" containername
You can access the host using its ip but localhost/127.0.0.1 won't work (they will resolve to the Linux VM that is part of docker for windows). If you use the default network settings, your host should be reachable on 10.0.75.1 from your container
I have a docker image and when I run it I need to pass the host machine IP address as an environment variable. So I need something like this:
docker run --rm -it -e HOST_IP=<?????> -p 8000:8000 image
I am using Docker on OS X. Basically this image is running a service that I want to connect to my local PostgreSQL server. The service reads the server host IP from an environment variable.
How do I get the IP of the host machine for docker to use?
If I use local machine address 192.168.99.1 (from ifconfig), psycopg2 complains:
psycopg2.OperationalError: FATAL: no pg_hba.conf entry for host "192.168.99.100", user "postgres", database "database", SSL off
The address 192.168.99.100 mentioned in the error is the IP of my docker-machine.
How can I get the correct IP?
I'm not sure you could visit the host IP inside the docker container. Because I believe they are inside different network.
Instead of visit pgSQL on host, you should run pgSQL in another docker container, then use docker-compose.yml to connect the two docker containers, so they could link to each other, and visit each other by IP.
More details on: https://docs.docker.com/compose/compose-file/