Ubuntu – PostgreSQL can’t bind to docker ip at boot (Cannot assign requested address)

dockerpostgresqlsystemdUbuntu

I am running PostgreSQL on the host machine. An application inside of a docker container should be able to communicate with the database. Since postgres is by default only listening for connections on localhost, I've changed the listen_addresses by adding the IP of the machine on the docker0 network interface, 172.17.0.1:

postgresql.conf:

[...]
listen_addresses = '127.0.0.1,172.17.0.1'
[...]

Changing the postgresql.conf and restarting the database works fine, it does what it should. However, when I'm rebooting the machine, I see that postgres isn't able to bind to the interface:

postgresql-10-main.log:

2019-12-09 23:03:10.202 UTC [935] LOG:  listening on IPv4 address "127.0.0.1", port 5432
2019-12-09 23:03:10.218 UTC [935] LOG:  could not bind IPv4 address "172.17.0.1": Cannot assign requested address
2019-12-09 23:03:10.218 UTC [935] HINT:  Is another postmaster already running on port 5432? If not, wait a few seconds and retry.
2019-12-09 23:03:10.218 UTC [935] WARNING:  could not create listen socket for "172.17.0.1"
[...]

Again, restarting postgres works fine – but I would like to not restart the database manually after a system reboot.

(Ubuntu 18.04, PostgreSQL installed via apt, Docker version 19.03.5, build 633a0ea838)

What I've tried

I was suspecting that the docker service or the network must start before postgres, so I tried adding a systemd file, but it doesn't work

/etc/systemd/system/postgresql.service.d/override.conf:

[Unit]
After=docker.service
After=network.target
Require=docker.service
Require=network.target

All the tutorials I can find suggest setting listen_addresses = '*', but since the server is directly connected to the internet I would like to avoid that.

Best Answer

In addition of overriding postgresql.service, override the postgresql@10-main.service as well.

For instance with systemctl edit postgresql@10-main.service then add:

[Unit]
After=docker.service

Related Solutions

Nginx failed to start. Cannot assign requested address

bind() to [ipv6]:80 failed (99: Cannot assign requested address)

It sounds like your IPv6 address has just been assigned to eth0, and still be in the tentative state, thus Nginx cannot establish a listen on that IP.

Either turn of DAD (Duplicate Address Detection) by running:

sysctl -w net.ipv6.conf.eth0.dad_transmits=0

(insert into /etc/sysctl.conf for permanent)

or add this line into /etc/network/interfaces, under inet6 interface definition

post-up while ip -6 addr show tentative | grep . > /dev/null ; do sleep 1 ; done

Source: http://pyro.eu.org/how-to/micro/nginx-cannot-assign-requested-address-ipv6.txt

Ubuntu – How to auto-start docker containers at system boot

Apparently, the current method to auto-start Docker containers (from Docker 1.2) is to use restart policies. This will control how Docker should handle starting of the container upon startup and re-starting of the container when it exits. I've used the 'always' option so far, and can confirm that it makes Docker auto-start the container at system boot:

sudo docker run --restart=always -d myimage

Documentation Excerpt

Restart Policies Using the --restart flag on Docker run you can specify a restart policy for how a container should or should not be restarted on exit.

no - Do not restart the container when it exits.

on-failure - Restart the container only if it exits with a non zero exit status.

always - Always restart the container regardless of the exit status.

You can also specify the maximum amount of times Docker will try to restart the container when using the on-failure policy. The default is that Docker will try forever to restart the container.
$ sudo docker run --restart=always redis
This will run the redis container with a restart policy of always so that if the container exits, Docker will restart it.
$ sudo docker run --restart=on-failure:10 redis
This will run the redis container with a restart policy of on-failure and a maximum restart count of 10. If the redis container exits with a non-zero exit status more than 10 times in a row Docker will abort trying to restart the container. Providing a maximum restart limit is only valid for the on-failure policy.