I’m trying to run Livebook locally on MacOSX. It was working on Friday and I’d guess the most likely cause of failure is some setting the corporate folks have changed. What I want to find out is if there’s any way I can gather more information on the failure.
I get a -1 when I attempt to start it. The following is in the log file:
[LivebookLauncher] release pid: 51007
ERROR!!! [Livebook] Your hostname “xxxxxxx” does not resolve to a loopback address (127.0.0.0/8),
which indicates something wrong in your OS configuration, or EPMD is not running.
Make sure your computer’s name resolves locally or start Livebook using a long distribution name. Please try one of the fixes below:
If you are using the Livebook App, please open up a bug report.
If you are using Livebook’s CLI, consider using longnames:
livebook server --name firstname.lastname@example.org
If you are running it from source, do instead:
MIX_ENV=prod elixir --name email@example.com -S mix phx.server
[os_mon] memory supervisor port (memsup): Erlang has closed
[os_mon] cpu supervisor port (cpu_sup): Erlang has closed
I removed my actual host name from the log (if you’re wondering why my hostname is “xxxxx”).
I tried manually starting EPMD. No error but it didn’t fix the issue either.
I’ve tried uninstalling and reinstalling Livebook and that doesn’t do it. I’ve done a bit of digging and in my case the host name does not resolve to 127.0.0.1 (which strikes me as kind of odd). This is why I think it’s some sort of corporation setting. I mean if I
ping hostname it doesn’t show me 127.0.0.1
Anyone have a notion of what I should be looking at to gather more detail on this problem?
How are you running Livebook? I ask because in the app I believe epmd is embedded so restarting a global version likely doesn’t do much.
Have you tried restarting or turning WiFi off and on again to try to see if you get a different IP or perhaps corp settings were reverted? If
ping hostname resolves as hex like 0000:0000:0000:0000 then it’s likely trying to resolve as the ipv6 version of localhost.
Thanks for trying to assist. Whatever it was must have been transient because today Livebook is working locally just fine. I shut off my machine and restarted it a couple of times yesterday but that didn’t fix it so I’m puzzled as to what the issue was but it’s working now.
Yeah that’s weird. Maybe corporate reverted whatever changes broke it? It definitely sounds network related but that’s a big bucket to troubleshoot on its own.
You could try running
ifconfig and store the output in a text file. That should tell you all the network interfaces and ip addresses associated for a working state. Then if/when this happens again you can run the command to compare what changed. For most networks, only the IP addresses change over time. You can run the command with arguments to focus on just localhost but there may be things you can spot on the other interfaces.
According to GitHub - livebook-dev/livebook: Interactive and collaborative code notebooks for Elixir - built with Phoenix LiveView there is a
LIVEBOOK_IP variable you can bind to. It mentions needing a valid ipv4 or ipv6 address but you could try binding it to your corporate network IP address next time. My guess is something about the address it detected didn’t seem valid for whatever reason. I don’t know how or if it’s possible to change these settings for the app. It may be possible to set it in a terminal window and then use
open /Applications/Livebook.app to affect the context the app uses. That may not work and require the CLI version to modify those env variables.
Thanks @w0rd-driven I encountered a similar issue with M2, and I was able to resolve it by simply restarting the Wi-Fi connection.