0

I have a small and simple UDP server that's been working in several hosting places for years.

I just created a new vm in the google cloud, a Compute Engine, with linux.

Default settings for a small machine (f1-micro (1 vCPU, 0.6 GB memory) CPU platform Intel Sandy Bridge), with only changes:

  • static public IP
  • inbound firewall rule to allow UDP traffic at the server port (22000)

Everything works great and can be connected from anywhere in the world, except from the machine itself (there's a stats module that asks the server for info - this also work from remote, but not from the machine itself). This worked in all other linux machines I tried the server on.

I was pointed at changing the address the server binds to as a possible solution, and as I have access to the source code I tried all I can imagine to no effect (binding to 127.0.0.1, 0.0.0.0, "localhost", specific IP...).

I have also tried adding a specific inbound and outbound firewall rules.

There are tons of information around about how to fix when you cannot connect from outside - my problem is the opposite, and I'm of ideas.

Ping etc work normally.

Thanks in advance

Improve this question
1
  • How is the server configured to listen? How is the client configured to connect? Commented Aug 3, 2018 at 15:20

1 Answer 1

Reset to default
0

Solved!

Apparently in these machines / kernel UDP sockets need to have a name (bind) before sending - never needed in other machines I installed the server.

Not a server administration thing - it just seemed as if it could be.

Improve this answer

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.