Closed theottm closed 5 years ago
Thanks for the report! The upstream BVG endpoint (that 1.bvg.transport.rest
wraps around) blocks the IP address of my server, so 1.bvg.transport.rest
returns "service unavailable" responses.
This can be solved by migrating the server to a new IP or using a different server. Let me know if you have one available.
I have a sleeping EC2 instance that could be used for that.
Tell me how I should have the service running.
I tried to run the docker container as described here with no success. The result message is the same. Never worked with node so I did not tried the npm solution yet, but I could.
I have a sleeping EC2 instance that could be used for that.
Cool!
Tell me how I should have the service running.
How would you like to run the service? Via Docker or manually via common Node.js tooling?
I tried to run the docker container as described here with no success. The result message is the same.
Does that mean you get the "no connection available" message back?
I've just tried and running locally via Docker works fine, I get data back. (I checked with curl 'http://localhost:3000/stations/900000100001/departures?duration=1'
.)
Seems like the service has recovered. Nevertheless, it's helpful to know a) that you're willing to host it and b) what problems you face with the setup.
Yes I did receive the same error message. But if it worked for you I will try to the Docker container again.
Oh great ! I will still try to set up the server, so we can use it in case of another problem :)
It's down again... maybe we should try a rotating IP solution with a proxy list or something.
For now I am not able to run the container properly on my local machine
curl 'localhost:3000/journeys?from=900000017104&to=900000017101'
while the container is running returns me :
{"error":true,"msg":"VBB error: Service Unavailable"}
VBB error: Service Unavailable
AFAIR that is their endpoint reporting downtime, not your IP being blocked. Errors like ECONNRESET
or ETIMEDOUT
hint that your IP has been blocked.
We investigated further and found a bug: The install-unique ID generated by hafas-client
(which is a dependency of bvg-rest
) is being built into the Docker image. It should be regenerated on container start in addition.
@theottm Would you mind submitting the fix?
Thanks, didn't find the time today. Glad it works again :)
1.bvg.transport.rest seems down