Can't reconnect to TOR nodes after migration

[miketwenty1]

Background

Migrated from 0.12.1 to commit hash 213b264 (master at the time)

There were a good many TOR nodes that were previously connected to our node.

Channels and peers with tor nodes didn't reconnect after migration, they remain inactive, we are using clear net only.

Is there a current strategy to reconnect with these nodes?

[miketwenty1]

@guggero is this possible?.. connecting with an existing tor peer from clearnet after a migration?

[wpaulino]

If you're on clear net only, you'll have to wait for Tor nodes to reconnect to you.

[andrerfneves]

@wpaulino the only way to force the Tor nodes to reconnect would be to lncli connect pubkey@newIP:9735 or by restarting their node, which will effectively perform checks on peers and channels on startup, etc, etc. Is this correct?

Is there any way to avoid this from happening in the future? If we were both tor and clear net, then this wouldn't be an issue I presume?

[wpaulino]

We have reconnection logic, regardless of what network you're running on, that runs after a peer disconnects, so a restart or manual reconnection aren't needed. There could be an issue with the reconnection logic though as noted in #5189. Are you able to get in contact with the other node and see why it's not reconnecting to you?

Is there any way to avoid this from happening in the future? If we were both tor and clear net, then this wouldn't be an issue I presume?

It would only help in that you are now also able to reconnect to them through Tor, but they should still be able to reconnect to you regardless. All of your traffic would also go through Tor until we have #5042.

[andrerfneves]

@wpaulino thanks for the information. We had two tor-only node peers reboot their nodes and the channels have since become active. That's why I figured the reconnection through lncli would perform the same function.

Can I fetch any logs/info from our node to see if it really is the issues pointed out in #5189?

[wpaulino]

We would need the remote node's logs to see why they aren't reconnecting (assuming you don't see any rejected reconnection attempts from them in your logs).

[Roasbeef]

This should have ben addressed by #5538 and #5794