danieldent / docker-postgres-replication Goto Github PK

I am a complete beginner in Postgres, and I have pulled your docker master/slave image but it seems to be lacking the documentation onto what should be done after this pull.
I have tried docker run danieldent/postgres-replication but it takes me to a series of log messages and takes away the shell to actually do anything, please guide me in this regard
Thanks

I'm migrating the docker-compose.yml to kubernetes, but I'm facing an issue that it seems to me has nothing to do with k8s, but with postgres, and I'm not an expert in postgres.

I got the following error when connecting a slave to a successfully deployed (and functional) master:

2016-10-26T10:39:48.274183765Z pg_basebackup: could not connect to server: FATAL:  role "-vP" does not exist
2016-10-26T10:39:48.274942692Z Waiting for master to connect...

The message repeats endlessly because it's inside a until/done structure.

The k8s pod file below:

apiVersion: v1
kind: Pod
metadata:
  name: postgres-slave
  labels:
    name: postgres-slave
spec:
  containers:
    - name: postgres-slave
      image: danieldent/postgres-replication:9.6
      env:
        - name: POSTGRES_USER
          value: "postgres"
        - name: POSTGRES_PASSWORD
          value: "postgres"
        - name: PGDATA
          value: "/var/lib/postgresql/data/pgdata"
        - name: REPLICATE_FROM
          value: "10.1.5.4"
      ports:
        - containerPort: 5432
      volumeMounts:
        - mountPath: /var/lib/postgresql/data
          name: pg-data-slave
  volumes:
    - name: pg-data-slave
      emptyDir: {}

The container pings successfully to the master in 10.1.5.4 (I know I should connect to a service name, but this is a test), but afterwards I got the error above.

Now, if I change POSTGRES_USER and POSTGRES_PASSWORD to the deprecated version (PGUSER and PGPASSWORD), then the container passes the endless loop and replicates from the master correctly.

Excuse my newb-ness on this (and if there's a better place to ask, let me know), but I cannot seem to actually connect to the database once this is running. I cloned the repo, didn't change anything (as I just need this for some local dev testing), and did docker-compose up. It seems to run fine, as I see this in the output:

pg-master_1  | LOG:  database system is ready to accept connections
pg-slave_1   | LOG:  database system is ready to accept read only connections

I see the two containers in docker ps, etc. I then tried to actually attach to it, e.g.:

psql -h 172.18.0.2 -U postgres -W

IP address I got from doing docker inspect on the master's container ID. This eventually just times out. I'm running this on a Mac. Could you clue me in to what I'm missing/doing wrong?

Hi thanks for your code,
i'm looking for a solution that don't use persistent volumes. My goal is to be able to reinitializate the postgres data inside the master container, recovering them from a replica that is still alive, without the needs of share a common folder.I don't want to mount anything inside the container and be sure that if 1 replica stay up, all the data will be restored on master and the other replicas. Can you help me?

Hi, I just want to know the current status of this project as the last commit is from more than a year ago. Would be a waste if this project ends up dying. And want to say that I am interested in helping to maintain this project. What would be the first steps to do so?

Hi!

I'm probably stupid but I can't connect to the database with laravel. What should I put in the .env file?
I've tried everything:

• port tested: 5432
• host tested: localhost, 127.0.0.0.1, pgsql

this is my docker-composer:

services:

  php:
    image: swithfr/php:7.2
    volumes:
      - ".:/var/www/laratest"

  nginx:
    image: nginx:latest
    volumes:
      - "./docker/nginx.conf:/etc/nginx/conf.d/default.conf"
    links:
      - php
    ports:
      - 8888:80

  pg-master:
      image: swithfr/pgsql-replication
      restart:'always'.
      environment:
        POSTGRES_USER: "postgres
        POSTGRES_PASSWORD:'postgres'.
        PGDATA:'/var/lib/postgresql/data/pgdata'
      volumes:
       - '/var/lib/postgresql/data'
      exposes:
       - '5432'

  pg-slave:
    image: swithfr/pgsql-replication
    restart:'always'.
    environment:
      POSTGRES_USER: "postgres
      POSTGRES_PASSWORD:'postgres'.
      PGDATA:'/var/lib/postgresql/data/pgdata'
      REPLICATE_FROM: "pg-master
    volumes:
     - '/var/lib/postgresql/data'
    exposes:
     - '5432'
    links:
     - "pg-master

swithfr/pgsql-replication is a copy of your Dockerfile with just the addition of these lines

RUN apt-get update 
RUN apt-get install iputils-ping -y

I had to copy-paste file_env() for shared secrets, and there are some more changes for overall progress.

Also, what is the original version of docker-entrypoint.sh you started from ? I could try to merge automatically, but it's impossible without complete tree and it's root.

pg-slave_1 | Waiting for master to ping...
pg-slave_1 | /usr/local/bin/docker-entrypoint.sh: line 39: ping: command not found

It is necessary to install ping. It is solved adding in the Dockerfile:

RUN apt-get update
RUN apt-get install iputils-ping -y

Awesome write up. Replication run fine in my side. Container data also works at '/var/lib/postgresql/data/pgdata'

Postgresql is installed in my local machine. But data dose not mount in below location

volumes:
- '/var/lib/postgresql/data'

Hey Daniel, any chance you could add a license to this project?

I am not very familiar with docker-compose

But this sample looks like running master & slave on the same machine.

If I am right, is there a way to run them on two different machines?