PostgreSQL Private Network Unreachable After Upgrade to Pro Plan
shwayyan
PROOP

3 months ago

I upgraded my project to the Pro plan and restarted my PostgreSQL service due to disk space issues. After the restart, my backend service can no

longer connect to PostgreSQL via the private network.

Error Details:

- Error: connect ENETUNREACH fd12:3b8f:a084:0:a000:42:8798:d8c9:5432

- Project: radiant-enthusiasm

- PostgreSQL service shows as "Running" in dashboard

- Public connections work (I can connect via mainline.proxy.rlwy.net:41925)

- Private IPv6 network connections fail with ENETUNREACH

What Happened:

1. Database ran out of disk space

2. Upgraded to Pro plan

3. Restarted PostgreSQL service

4. PostgreSQL got new IPv6 address: fd12:3b8f:a084:0:a000:42:8798:d8c9

5. Backend service cannot reach this new IPv6 address on private network

6. Multiple backend redeploys haven't resolved the issue

Current State:

- PostgreSQL: Running but unreachable via private network

- Backend: Crash loop due to ENETUNREACH

- Both services are in the same project/environment

Could you please help reset the private network connectivity between these services?

Project ID: 42c1a6ca-79d1-4577-9cc1-448d5819c849

Thank you!

$10 Bounty

3 Replies

Railway
BOT

3 months ago

Hey there! We've found the following might help you get unblocked faster:

If you find the answer from one of these, please let us know by solving the thread!

brody
EMPLOYEE

3 months ago

This thread has been marked as public for community involvement, as it does not contain any sensitive or personal information. Any further activity in this thread will be visible to everyone.

Status changed to Open brody • 3 months ago

mykal
PRO

3 months ago

Hey, a few questions / tips that I had when reading through this:

  • Are you using the IPv6 address directly in your backend? I would recommend you stick to the private networking DNS names. Just make sure your backend is making calls via the postgres.railway.internal DNS name, not the IPv6 address directly. IP addresses on railway are ephemeral between builds afaik and will change when you redeploy your service.

  • Is your backend service failing during the build? or in the deployment step? if it's during the build, that's probably because Railway doesn't support the use of private networking during the build step due to how their builders work (they don't live in the same virtual env as your project yet). If it's during your build, make sure to use the public connection string for builds and the private/internal connection string for when the actual app runs.

After doing some messing around, and reading through the changelog, the railway docs are out of date and private networking does actually support IPv4 now as well. That said, if you suspect the issue has to do with IPv6 you can move back to IPv4 happily by just grabbing the IPv4 address of the service.

gstreetmedia
PRO

3 months ago

I'm having this same issue. I cannot connect my node application to either postgres or redis using private networking. It connects to the public URL's just fine during build. I did a little exploration using SSH + nslookup and neither of the domains can be found

root@4c1522e95c57:/app# nslookup -query=AAAA postgres.railway.internal

;; Got recursion not available from fd12::10

Server: fd12::10

Address: fd12::10#53

** server can't find postgres.railway.internal: NXDOMAIN

root@4c1522e95c57:/app# nslookup -query=AAAA redis.railway.internal

;; Got recursion not available from fd12::10

Server: fd12::10

Address: fd12::10#53

** server can't find redis.railway.internal: NXDOMAIN

Loading...