Forced disconnection every 24 hours by my ISP

Node Operators
, ,
1

Hello everyone,

I live in Germany and my ISP disconnects my connection every 24 hours. This reduces my online time since my node only starts receiving data again after an hour. I have already looked at the config.yaml to see which parameter I need to change, but I haven’t come to any conclusion. Does anyone know the correct parameter for me? I have a Windows Server. My plan for 2026 is to register a small business, then I can also change the connection and will have a fixed IPv4. Unfortunately, there is no fixed IP for private customers. Thank you.

2

How long is the disconnection to isp?

Does your external IP address change?

How is DDNS update done?

Are you sure it’s ISP disconnecting you?

A short interruption should not effect storjnode at all, less than 5min and your unlucky if it hurts online score.

1 Like
3

I have the relevant information. Since the node only resolves the name at startup and does not do so at intervals, the only option I have is to restart the node every 24 hours. But what I don’t understand is why exactly after one hour.

4

The solution is to use some kind of dynamic dns service. Your router likely has support for such services built in. What router is it?

1 Like
5

Then tell us.

The node doesn’t resolve its own name, the satellite and the customers do.

So do you have DynDNS set up? If yes, which one? Usually DynDNS providers use a TTL of 5 minutes or even less.

Do you have a Fritzbox? Then you could use their own service Myfritz as DynDNS.

2 Likes
6

ISP should provide a DynDNS service for personal plan. Ask them.

1 Like
7

I’ve seen the same behavior from the SAT that when a nodes IP changes via DNS, it usually takes up to 1 hours for all clients to use the new IP. My guess is that SAT does not resolve it very often ?

However, if the node restarts and connects back to the SAT with new information, it is updated instantly. (by design i don’t know)

IF this in fact is your ISP forcing you to a new public IP every 24 hours, one solution could be to make a script to monitor this behavior, when observed, change the IP in the node configuration and restart it. This should make the downtime from data shorter.

However - it does sound strange that an ISP would force such behavior, and I would persue other solutions first and this is not nice :slight_smile:

8

Most german ISPs doing this. They disconnect every 24 hour even if IP doesn’t change.

I don’t know the reason, maybe just to keep fixed IP as feature they can sell at a higher price.

9

Yes, in Germany it is normal for the ISP to disconnect VDSL connections every 24 hours. You then receive a new public IPv4 address every day. Unfortunately, they do not offer a fixed IP for private customers, only for business customers. Otherwise, I would have done it a long time ago. Here in our country, we have only about 14% fiber optic connections. The rest is cable or DSL. Poor Germany

10

Once again, the solution to your non-problem is DynDNS.

1 Like
11

Maybe they just restart the routers/gateways once per day.
I believe in most european countries it’s the same, for personal plans you get dinamic IP.
But, at least one of my ISPs offer free DynDNS. So, dinamic IP is not a problem. They don’t disconnect me once per day, but I scheduled my router to restart once per day, just because… The connection is recovered quick, though, in less than 5 min.

12

The satellite has the cache, however, if it cannot contact the node (which is initiated by the node on every check-in, every hour by default), the satellite will resolve it again and will update the cache, if your node is offered to the customer.

This is exactly what’s happened - the node will try to check-in on the satellite and the satellite will resolve the provided external address to try to contact your node.

It’s likely a TTL of your DNS service. You may configure it to a more short interval or use DDNS service like NoIP, which already has a short TTL.

13

This attitude is my solution.

before

#how frequently the node contact chore should run
#contact.interval: 1h0m0s

after

#how frequently the node contact chore should run
contact.interval: 0h5m0s

It’s going wonderfully.

1 Like
14

I’ve tested with TTL as low as 1minute - still about 1 hour before the SAT picks up the changes.

15

Then I would assume, that the domain is updated only on the node’s check-in.

16

At Telekom, for example, there has been no daily forced disconnection for a long time. A session can last for around six months without interruption. However, you can still set up a daily disconnection in your router.

17

First, check your equipment. Short-term outages are usually a client-side issue. If you’re using FTTX, upgrade to a more powerful router. If you’re using GPON, upgrade to a more powerful terminal. As a last resort, check the equipment’s temperature over long periods of operation and install additional cooling. Simple consumer equipment on the client side isn’t designed to handle heavy traffic over long periods of time.