2 years ago
48f3719a-323b-4609-812c-5b818714b649 deployment id, 2d76fc7b-e09c-410f-b7e3-117dfccc41bf project id.
I have a service with private networking on that should be available on personalized_assistant_core.railway.internal
My other deployed service can't seem to connect to it over internal netwok, neither by internal url : Could not connect to personalized_assistant_core.railway.internal:8786. Error: Timed out trying to connect to tcp://personalized_assistant_core.railway.internal:8786 after 30 s
nor by the resolved url : OSError: Timed out trying to connect to tcp://[fd12:ce2b:d076:0:4000:4:e7f5:36d8]:8786 after 30 s
I deploy using dockerfile FROM python:3.10, I tried setting env var ENABLE_ALPINE_PRIVATE_NETWORKING=true.
What could be the root of the problem? personalized_assistant_core.railway.internal does have one port exposed to the web, but i'm connecting to the other port, can that be an issue?
4 Replies
Actually, now that I think about it, this specific port might be not exposed to internal railway network, the dask cluster I want to connect to (personalized_assistant_core service) listens at tcp://127.0.0.1:8786. 🤔 Is it possible to somehow config it to be exposed/binded properly?
Ok, the docs (You will need to bind to a IPv6 port to receive traffic on the private network) and duck debugging help.
Ty, dear duck, you're extremely useful!
<:duck50:755227796414136410>
Problem resolved 🌚
2 years ago
fwiw there's no such thing as exposed or not exposed on the private network, as there is no firewalls or anything of the sort on the private network, as long as it listens on ipv6 it doesn't matter the port or the protocol it will work.
fyi python:3.10 is not an alpine based image, it's based on Debian.