Node disconnects every 24 hours

Hello,

I can’t find a permanent solution to this problem. Here is what happens:

It works fine, the ports are open (verified by portchecker), no error messages, nodes communicating correctly.

After what seems to be about 24 hours, I always get this same error message.

Relay log:
[relay:cardano.node.IpSubscription:Error:4810] [2021-06-17 00:50:09.34 UTC] IPs: 0.0.0.0:0 [24.91.139.119:3001] Connection Attempt Exception, destination 24.91.139.119:3001 exception: Network.Socket.connect: <socket: 55>: does not exist (No route to host)

[relay:cardano.node.IpSubscription:Error:306] [2021-06-17 00:50:26.29 UTC] IPs: 0.0.0.0:0 [24.91.139.119:3001] Failed to start all required subscriptions

BP log:
[bpnode:cardano.node.IpSubscription:Error:4167] [2021-06-17 00:50:10.46 UTC] IPs: 0.0.0.0:0 [24.91.139.119:3002] Connection Attempt Exception, destination 24.91.139.119:3002 exception: Network.Socket.connect: <socket: 42>: does not exist (No route to host)

[bpnode:cardano.node.IpSubscription:Error:306] [2021-06-17 00:50:10.46 UTC] IPs: 0.0.0.0:0 [24.91.139.119:3002] Failed to start all required subscriptions

Port checker will then show that both ports are now closed, thus meaning the node is not connected.

To temporarily solve this disconnect…all I do is go to ‘settings’ and click the intel ethernet from on, to off, to back on, for both nodes.

After that, the log will give this warning sign:

Relay:
[relay:cardano.node.ErrorPolicy:Warning:314] [2021-06-17 00:50:41.74 UTC] IP 45.33.50.110:45178 ErrorPolicySuspendPeer (Just (ApplicationExceptionTrace (MuxError MuxBearerClosed “<socket: 29> closed when reading data, waiting on next header True”))) 20s 20s

Bp:
[bpnode:cardano.node.ErrorPolicy:Warning:77] [2021-06-17 00:50:41.74 UTC] IP 45.33.50.110:48142 ErrorPolicySuspendPeer (Just (ApplicationExceptionTrace (MuxError MuxBearerClosed “<socket: 36> closed when reading data, waiting on next header True”))) 20s 20s

After this, the log runs perfectly and when I check port checker, both ports are open again.
Then, everything is good and the node runs for another 24 hours, before encountering the same issue.

Any help is greatly appreciated.

Can be the topologyupdater which is restarting the relays each 24 hours by default?

  • cnode-tu-restart.timer : schedules the cardano-node restart service, default every 24h

u can check the status with:

sudo systemctl status cnode-tu-restart.timer

1 Like

Hello - this is what I get:

Unit cnode-tu-restart.timer could not be found.

This is normal for a relay to update topology. My relay closes ports then opens them every 24 hrs. I believe the idea is to keep connections fresh. Not sure really… lol