These are mostly questions for n8n, not railway, but yes, you can recover the data.

Go to your postgres db -> settings -> networking -> create a TCP proxy pointed to port 5432

Then, go to the "data" tab. You will be able to view the contents of the database from there

I have this message in the screenshot, but when I check my variables they are correctly filled in.

## DNS Resolution Failure in Railway Shell Preventing PostgreSQL Connection (Private Endpoints) - TCP Proxy Created, Data Tab Empty

Dear Railway.com Support Team,

I am encountering persistent issues connecting to my PostgreSQL database from the railway shell in my project. I have already gone through extensive troubleshooting steps and would greatly appreciate your assistance in resolving what appears to be a DNS resolution problem.

Initially, the environment variables $PGHOST and $PGPORT were empty within the railway shell, preventing any connection. Following previous support suggestions, I created a TCP Proxy for the PostgreSQL service. While this did help populate the $PGHOST and $PGPORT variables, as per your previous suggestion, I also checked the "Data" tab in my PostgreSQL service after creating the TCP Proxy, but I am still unable to visualize the database content there.

More recently, I switched the variables to utilize private endpoints, and $PGHOST and $PGPORT now return the correct private endpoint values:

• echo "$PGHOST" returns: postgres.railway.internal

• echo "$PGPORT" returns: 5432

However, the connection to the PostgreSQL database still fails. The error I receive when attempting to connect using psql "$DATABASE_URL" or psql -h "$PGHOST" -p "$PGPORT" -U "$POSTGRES_USER" -d "$POSTGRES_DB" is as follows:

      psql: error: could not translate host name "postgres.railway.internal" to address: nodename nor servname provided, or not known
    

To investigate further, I executed the CRITICAL command ping postgres.railway.internal INSIDE the railway shell. The result of the ping was:

      ping: cannot resolve postgres.railway.internal: Unknown host
    

This ping result confirms that the issue is a DNS RESOLUTION FAILURE for PRIVATE hostnames (like postgres.railway.internal) WITHIN my railway shell environment. It appears the railway shell is unable to resolve the IP address corresponding to the private hostname postgres.railway.internal.

The $RAILWAY_PRIVATE_DOMAIN variable is correctly configured and returns the expected value:

• echo "$RAILWAY_PRIVATE_DOMAIN" returns: postgres.railway.internal

I'd like to request your assistance in resolving this DNS resolution issue, as it is also preventing me from visualizing my database content in the "Data" tab, even after creating the TCP Proxy as previously suggested.

Please find below my project and PostgreSQL service information to facilitate your investigation:

• Project Name: n8n-railway

• PostgreSQL Service ID: 153e57e9-7b45-49f4-b604-ce0c9fe06ece

• Environment: Production

Thank you very much for your attention and help in resolving this DNS resolution problem as quickly as possible, and for any guidance on accessing my database content in the "Data" tab.

Sincerely,

Carlos Guimarães

You're trying to ping the railway internal network which only works for our services. You'd want to use the TCP proxy domain/port to connect to the database. If you want to access the internal railway network, you can look into setting up tailscale: https://docs.railway.com/tutorials/set-up-a-tailscale-subnet-router

Otherwise, ensure the variables for PGHOST and PGPORT are set to the tcp proxy.

Hi Team,

Thank you for suggesting the Tailscale Subnet Router to access the PostgreSQL database and recover my n8n workflows.

I followed the instructions and the official "Tailscale Subnet Router" template from Railway.com, completing these steps:

1. Generated a Tailscale Auth Key.

2. Configured Split DNS in Tailscale as per documentation.

3. Deployed the "Tailscale Subnet Router" template in my project, inserting the Auth Key.

4. Attempted deployment with the default template "Source image": tailscale/tailscale:stable.

5. Due to failures, I tried changing the "Source image" to tailscale/tailscale:latest and tailscale/tailscale:v1.20.1.

6. In all attempts, the "Tailscale" service deployment failed with the persistent error: "This image does not have a tag."

7. No deployment logs are generated, only the message: "No deployment logs. All output sent to stdout will be displayed here".

This issue persists regardless of the Docker image tag used. I suspect the failure is on the Railway.com platform's side when attempting to pull the tailscale/tailscale image from Docker Hub.

Given the persistent error and lack of detailed logs, I await investigation from the Railway.com support team to identify and resolve this platform issue.

Sincerely,

Carlos Guimarães

Hi, we're investigating an issue with image pulls in metal regions: https://status.railway.com/cm8fwtzac0018ksgsby0yuaz7