Panic on my server (attempt to subtract with overflow)

[goomario]

🐛 Bug Report

Logs:

2021-12-09T12:10:08.758989Z  INFO Ledger successfully advanced to block 24433 (ab1xqkmfag68kuzpnl4wnkxlup2cvwf0dmxe7vxu9rrs8k30acmyczsjjy6sp)                                  
2021-12-09T12:10:12.591517Z  INFO Ledger successfully advanced to block 24434 (ab1gkjy5nsf7me9la7yaflfdgkgy0ted6k49j7m5g4g7zfusl8guc9qh9xk2r)                                  
2021-12-09T12:10:18.041998Z  INFO Ledger successfully advanced to block 24435 (ab1fp36u0eukug0s2g5tch93rt06zvvd2d7pymp57pav9fnamy3kyxqsxhka3)                                  
2021-12-09T12:10:18.167638Z  INFO Ledger successfully advanced to block 24436 (ab1v24ewezdz3xyhhd78e0g9qcxfgthv674z64fwa882gf2453r4sqqaxjt7u)                                  
thread 'tokio-runtime-worker' panicked at 'attempt to subtract with overflow', /aleo/snarkOS/src/network/ledger.rs:954:58                                 
stack backtrace:                                                                                                                                                               
   0: rust_begin_unwind                                                                                                                                                        
             at /rustc/f1edd0429582dd29cccacaf50fd134b05593bd9c/library/std/src/panicking.rs:517:5                                                                             
   1: core::panicking::panic_fmt                                                                                                                                               
             at /rustc/f1edd0429582dd29cccacaf50fd134b05593bd9c/library/core/src/panicking.rs:100:14                                                                           
   2: core::panicking::panic                                                                                                                                                   
             at /rustc/f1edd0429582dd29cccacaf50fd134b05593bd9c/library/core/src/panicking.rs:50:5                                                                             
   3: snarkos::network::ledger::Ledger<N,E>::update_block_requests::{{closure}}                                                                                                
             at ./src/network/ledger.rs:954:58
   4: <core::future::from_generator::GenFuture<T> as core::future::future::Future>::poll
             at /rustc/f1edd0429582dd29cccacaf50fd134b05593bd9c/library/core/src/future/mod.rs:80:19
   5: snarkos::network::ledger::Ledger<N,E>::update::{{closure}}
             at ./src/network/ledger.rs:295:17
   6: <core::future::from_generator::GenFuture<T> as core::future::future::Future>::poll
             at /rustc/f1edd0429582dd29cccacaf50fd134b05593bd9c/library/core/src/future/mod.rs:80:19
   7: snarkos::network::ledger::Ledger<N,E>::open::{{closure}}::{{closure}}
             at ./src/network/ledger.rs:200:21
   8: <core::future::from_generator::GenFuture<T> as core::future::future::Future>::poll
             at /rustc/f1edd0429582dd29cccacaf50fd134b05593bd9c/library/core/src/future/mod.rs:80:19
...
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace. 
2021-12-09T12:10:44.342685Z  WARN [UnconfirmedBlock] channel closed
2021-12-09T12:10:45.142860Z  WARN [UnconfirmedBlock] channel closed
2021-12-09T12:10:47.222083Z ERROR Failed to send heartbeat to ledger: channel closed
2021-12-09T12:10:50.706777Z  WARN [Pong] channel closed
2021-12-09T12:10:50.706905Z  WARN [Pong] channel closed
2021-12-09T12:10:50.707119Z  WARN [Pong] channel closed
2021-12-09T12:10:51.759396Z  WARN [Pong] channel closed
2021-12-09T12:10:51.761376Z  WARN [Pong] channel closed

Steps to Reproduce

Run by snarkos --trial --verbosity 0 --miner aleoxxxx

Your Environment

• testnet2 at 464586779a095a45158fb565a3c39627bb74956f
• rustc 1.57.0 (f1edd0429 2021-11-29)
• Ubuntu 18.04

[ljedrz]

While it's in line 955 (and not 954) in current testnet2 branch, I'm suspecting that it's the following operation:

let number_of_block_requests = std::cmp::min(maximum_block_height - latest_common_ancestor, E::MAXIMUM_BLOCK_REQUEST);

[ljedrz]

Upon closer inspection it seems that it's due to the current logic assuming that a greater combined weight means that the corresponding height is guaranteed to be greater.

[goomario]

While it's in line 955 (and not 954) in current testnet2 branch, I'm suspecting that it's the following operation:

let number_of_block_requests = std::cmp::min(maximum_block_height - latest_common_ancestor, E::MAXIMUM_BLOCK_REQUEST);

You are right! I'm looking at the code, maybe with add comment.

[ljedrz]

"Duct-taping" this with a saturating_sub is an option, but it would only conceal the underlying issue; there's a related TODO by @howardwu there, Ccing him so he can comment on this when he's available.

[ljedrz]

I don't think it applies to testnet3 anymore, so closing.