localtunnel: When connection crashes, can no longer use that subdomain

I understand that good work is being done on this library to help improve reliability. I wanted to bring up another issue that may be related in an adjacent way. I’m also curious if anyone else is having this issue.

My connection frequently disconnects, but I then can’t use that same subdomain… I don’t mind having to restart localtunnel every now and then, but losing access to that subdomain is pretty frustrating.

For example, I use mydomain as the subdomain on port 3000, it crashes, and when I run lt --subdomain mydomain --port 3000 again, it generates the random characters. I then have to use something like mydomain1, and update the server address on every single sensor I have (which send data over a cellular connection to my rails server), which is definitely frustrating.

Has anyone else had this issue?

Is there any way to kill all instances of localtunnel if there are instances running…?

About this issue

Original URL
State: closed
Created 6 years ago
Reactions: 26
Comments: 27 (5 by maintainers)

Most upvoted comments

I’ve deployed a new server backend that should resolve this issue. If your client is disconnected (or dies) then the subdomain will be available again within a second.

defunctzombie on May 16, 2018

@defunctzombie I use nodemon to automatically restart my node server and, sometimes, it also happens when I make some changes in my application.

Localtunnel logs: /usr/local/lib/node_modules/localtunnel/bin/client:58 throw err; ^ Error: connection refused: localtunnel.me:36312 (check your firewall settings) at Socket.<anonymous> (/usr/local/lib/node_modules/localtunnel/lib/TunnelCluster.js:47:32) at Socket.emit (events.js:127:13) at emitErrorNT (internal/streams/destroy.js:64:8) at process._tickCallback (internal/process/next_tick.js:114:19)

Then, I try to restart localtunnel but it gives me a different URL and ignores my subdomain.

saeta-eth on Apr 4, 2018

I am facing the same issue even after the patch.

abhinaykumar on May 29, 2018

Yep, same here. Would be great to have a --close flag to use on the command line…

jpsear on Feb 23, 2018

Just blocked two sub-domains again. Occurs more frequently when I am on VPN trying to access network resources in parallel. Suspect a lingering connection after exiting application (ctrl +c).

Easiest way to reproduce:

Open a tunnel lt -s <custom_subdomain> --port <port>
Once you receive desired sub-domain, disconnect wi-fi/ network adapter (to emulate possible connection drops)
Exit application (ctrl + c)
Connect back wi-fi/ network adapter
Attempt to open tunnel with same sub-domain lt -s <custom_subdomain> --port <port>
You will no longer be able to claim requested sub-domain

Failed to locate any mechanism to forcefully emit a reset/ dead or close request to trigger sock.destroy();. I suppose its back to you @defunctzombie .

CyberQuestor on Jun 19, 2018

When I said close the connection, I meant to say close the program in the terminal via pressing ctrl-c. The subdomain goes unavailable for at least 16+ hours (I haven’t tested it very rigorously).

RoboticMind on Apr 3, 2018

Also for the Reproducing the bug:

keep a connection open for a very long time and every hour or so, ping it and see if it times out
when the connection times out it won’t start work working again so close the connection
try to get the same subdomain again and you won’t be able to

RoboticMind on Apr 2, 2018

I have the same issue.

highfeed on Feb 26, 2018