Panic in UPnP search

[deffrian]

Description

Lighthouse crashes during startup

Version

a53830f (current master)

Present Behaviour

lighthouse bn --network sepolia --execution-endpoint http://localhost:8551 --execution-jwt <path> --checkpoint-sync-url https://sepolia.checkpoint-sync.ethpandaops.io --disable-deposit-contract-sync

Apr 06 13:19:14.045 INFO Execution payloads are pruned           service: freezer_db
thread 'tokio-runtime-worker' panicked at 'invalid url or method: Error(InvalidBaseUrl)', /home/deffrian/.cargo/registry/src/git.luolix.top-1ecc6299db9ec823/attohttpc-0.10.1/src/request.rs:94:41
stack backtrace:
   0: rust_begin_unwind
             at /rustc/2c8cc343237b8f7d5a3c3703e3a87f2eb2c54a74/library/std/src/panicking.rs:575:5
   1: core::panicking::panic_fmt
             at /rustc/2c8cc343237b8f7d5a3c3703e3a87f2eb2c54a74/library/core/src/panicking.rs:64:14
   2: core::result::unwrap_failed
             at /rustc/2c8cc343237b8f7d5a3c3703e3a87f2eb2c54a74/library/core/src/result.rs:1790:5
   3: attohttpc::request::RequestBuilder::new
   4: igd::search::search_gateway
   5: network::nat::construct_upnp_mappings
   6: tokio::loom::std::unsafe_cell::UnsafeCell<T>::with_mut
   7: tokio::runtime::task::core::Core<T,S>::poll
   8: tokio::runtime::task::harness::Harness<T,S>::poll
   9: tokio::runtime::blocking::pool::Inner::run
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.
Apr 06 13:19:14.066 CRIT Task panic. This is a bug!              advice: Please check above for a backtrace and notify the developers, message: <none>, task_name: UPnP
Apr 06 13:19:14.066 INFO Internal shutdown received              reason: Panic (fatal error)
Apr 06 13:19:14.066 INFO Shutting down..                         reason: Failure("Panic (fatal error)")
Apr 06 13:19:14.067 INFO Saved DHT state                         service: network
Apr 06 13:19:14.067 INFO Network service shutdown                service: network
Apr 06 13:19:14.067 INFO HTTP server is disabled
Apr 06 13:19:14.068 INFO Saved beacon chain to disk              service: beacon
Panic (fatal error)

Expected Behaviour

No exception

[deffrian]

also on 4d17fb3

[michaelsproul]

Which OS? Any virtualisation or containerisation in play? Looks like a weird error from the OS being unable to do UPnP.

You can probably work around it with --disable-upnp

[deffrian]

Thanks, works with --disable-upnp. Os: Ubuntu 22.04.2

[deffrian]

No virtualization or containers

[michaelsproul]

Our UPnP support is best-effort due to the vast number of routers and networking setups that are possible, but I think we could avoid panicking in this case (a clean failure would be better).

Which router do you have, if you don't mind me asking? You can also DM the answer on Discord, I'm @sproul#3907, you can cross-check my profile on the Lighthouse Discord.

[michaelsproul]

In terms of a fix, looks like maybe we'd have to go to the igd dep.

[deffrian]

I don't know it was on public wi-fi. I will try to ask owner of router in few days

[ackintosh]

I had a look at the igd a bit, perhaps the PR below might improve this issue:

sbstp/rust-igd#55

[ackintosh]

I have filed a PR to rust-igd. sbstp/rust-igd#56

[deffrian]

Sorry, I can't reproduce it anymore. Just prints
INFO UPnP not available error: IO error: Resource temporarily unavailable (os error 11), service: UPnP
On commit 4d17fb3

[ackintosh]

In current unstable branch, we have updated the dependency causing the panic. Now, if there's any issue while searching for a gateway, lighthouse will emit logs indicating what is wrong.

[divagant-martian]

Closed by #4257