connecting to observer on remote machine

Questions / Help
1 
Attempting to connect to quiet-cherry-7730@fdaa:0:3598:a7b:14bf:782c:be54:2
Erlang/OTP 24 [erts-12.2] [source] [64-bit] [smp:4:4] [ds:4:4:10] [async-threads:1] [jit]

Node Connected?: false
Connected Nodes: []

this is the output of bash ./observer when following this guide.

Why is Node Connected false?

I suspect because of that, I cannot see the remote node in my Observer.

1 Like
2

Hello @abeeshake!

Is this still an issue for you?

If so, here’s some of the points to check:

  • Are you setting the OTP cookie explicitly?
  • Is the server instance able to cluster with other server instances? (does clustering work? Because the observer connection works through clustering)
  • Do you have an active wireguard connection open? If it’s not open, then it can’t reach the instances running on Fly.
3

I haven’t been able to connect to the remote node either. For a start, it seems the script from the guide is broken due to changes to the fly CLI.

With the latest fly CLI it returns this error:
Error: unknown flag: --name

That’s probably because fly info has been deprecated.

I’ve tried to hardcode the values for my app_name and my IP address, but even with the WireGuard tunnel active, it never connects to the remote node.

Maybe an update to the guide and the script would be helpful?

Thank you!

1 Like
4

try fly status -a app-name

5

Hi,
I’m also struggling a bit with this after following the guides:

  • The static cookie is set up correctly.
  • Wireguard is set up correctly and the dig command is able to ping the Fly.io account and bring the application names.
  • The Node.connect() function just won’t connect (false return) even hardcoding application name and private/public IPV6 addresses, nothing seems to work. Has anyone been able to connect remotely?

Thanks.

1 Like
6

Hi… The format for FULL_NODE_NAME changed relatively recently, to…

<APP-NAME>-<RELEASE-TAG>@<IP-ADDRESS>

And it looks like the guide is still using the older style.

I’d suggest fetching it directly from within iex:

$ fly ssh console
# /app/bin/app_name remote
> Node.self
:"app-name-01HZGUK389AO3J23KK7@fdaa:1:1234:abcd:efgh:effe:4044:2"
          ^^^^^^^^^^^^^^^^^^^^
                    |
previous way would have omitted this part

You may also need to change --sname my_remote to --name rem@<local-ipv6> (without the s after the hyphens), where the latter is an address that the remote machine can still access.

Hope this helps a little!

7

I’m facing the same issue, even with the correct full node name. The tunnel is setup and running, but I get the following every time:

Node Connected?: false
Connected Nodes: []
Erlang/OTP 26 [erts-14.2.5] [source] [64-bit] [smp:8:8] [ds:8:8:10] [async-threads:1] [jit]
8

Hm… The guide was updated in the interim, so I’m surprised that it’s still causing trouble…

Can you ping the fdaa:* address mentioned in the Attempting to connect message?

9

Hi @mayailurus :wave:,
If I run the following command:

ping fdaa:0:bd5c:a7b:1c3:****:****:*
ping: cannot resolve fdaa:0:bd5c:a7b:1c3:***:****:*: Unknown host

I get an unknown host error. The following command shows the apps though:

dig +noall +answer _apps.internal txt

Does it ring any bell? Thanks a lot :slight_smile:

10

Try ping6 or ping -6, instead. (Some implementations of ping need additional nudging for IPv6, if I recall correctly.)