Discussion Elite Dangerous Data Network (EDDN)

A new thread for this, mostly to help news get out about th emigration and slight downtime.

What is EDDN ?

The Elite: Dangerous Data Network is a system for willing Commanders to share dynamic data about the galaxy with others. By pooling data in a common format, tools and analyses can be produced that add an even greater depth and vibrancy to the in-game universe.

EDDN is not run by or affiliated with Frontier Developments.

Hosting is now provided by the EDCD community.

------------------------------------------------------------------------------------------------------

Heads up for some downtime on the EDDN service this coming Thursday 15th July 2021, around 07:00 UTC (i.e. during the game weekly maintenance).

We're moving the EDDN service to a new server in preparation for future improvement work.

We'll be leaving a 'bouncer' (to get messages to the new Gateway) and relay (to receive messages from the new Relay and send them on) up on the old host for as long as it still runs.

So, apart from a few minutes downtime whilst the old host is reconfigured and the new host processes confirmed up and running, there should be no interruption to service.

We've already reduced the DNS TTL for the eddn.edcd.io A record to 5 minutes, and it will be changed over to point at the new host ASAP Thursday morning once the configuration changes have been made.

If you're sending messages to EDDN using EDMC, EDD, EDDI etc then you might see a few 'connection refused' messages if it's still trying to send during the game downtime.

For people listening to the EDDN Relay output you should ensure that you restart your service after the downtime in order to pick up the new IP (135.181.213.148) rather than the old one (151.80.235.144). We advise manually checking how dig eddn.edcd.io A or similar resolves to ensure you're using the new IP. The combination of Bouncer and Relay left on the old host will ensure you still receive messages so long as the old host is still running, but you should verify that you're running directly against the new host ASAP.
 
Last edited:
The EDDN server migration is mostly now complete.

  1. The old host is now running a 'Bouncer' to forward all uploaded messages to the new host's Gateway.
  2. The old host is now running its Relay (what listeners connect to) such that it picks up messages from the new host Relay.
  3. The DNS for eddn.edcd.io has been changed over to point at the new host, 135.181.213.148 .
So, all uploaders will now be gradually migrating over to connecting to the new host Gateway, as they pick up the DNS change. Any still seeing the old IP will have their messages forwarded from old host to new host.

Any listeners that see the new DNS and have restarted will be receiving all messages directly from the new-host Relay.

Any listeners who have not yet reconnected, possibly because they're not yet seeing the new IP, will be receiving all messages from the reconfigured old-host Relay.

So, listeners!!! Check that the DNS has updated on your host, and if so restart your listener process(es) !

Once I'm sure the 'major' listeners (EDDN, Inara, EDSM, Spansh) are on the new Relay I'll begin periodically restarting the old host Relay in order to encourage more of the other listeners on to the new host directly.
 
...one of the 'things' that many of us take for granted (or that is completely invisible), but that is integral to the extended functionality and to the depth of the game. Thank you to the many who make this happen!
 
I restarted the old-host Relay this morning and that seems to have bumped most other listeners (I already knew the 'major' ones had moved) over to the new-host Relay. Hopefully the others will move over when I do the same for each of the next two days. After that I'll set something up to not only periodically restart, but also have the old-host Relay down for a short span of time when it does. If people start seeing 'Connection Refused' then maybe they'll think to ask what's going on and get the hint that their DNS cache needs flushing.

I can also see no new uploads to the old-host (going through the Bouncer to the new-host Gateway) since about 5 hours ago, and will continue monitoring that.

Hopefully these laggards are all just due to the DNS change taking a while to propogate (despite us having had the TTL at 5 minutes since Monday).
 
As it's now over 48 hours since the eddn.edcd.io A record was updated to the new IP, and we'd had the TTL on that record at 5 minutes for over 48 hours by then I did another restart of the old-host relay this morning. 6 listeners immediately reconnected to that old IP.

So, I've now set up a restart of the old-host relay for every 3 hours to see if that helps these laggards realise they should be using the new IP.

Tomorrow I'll start introducing some actual downtime into those restarts, gradually increasing. So, if anyone sees someone wondering why they're getting Connection Refused for an EDDN listener tell them to check this thread and flush their DNS cache.
 
I've now implemented that "some actual downtime" into the old-host Relay. It's starting at 1-2 minutes today and will gradually increase day on day.

So if you have an EDDN listener and are seeing disconnections once an hour, with connection refused for a while when that happens, you need to flush your DNS and actually start using the new-host Relay.
 
As per https://github.com/EDCD/EDDN/pull/120/files there is an upcoming change to allow the CodexEntry event through EDDN. Pay attention to the extra disallowed fields associated with this event before trying to send them.

That updated schema is now live at the reference URL but not yet actually in use on the Gateway. That will change after the during-game-maintenance restart of EDDN services tomorrow morning.
 
Top Bottom