Unable to trace an error from the backtrace caused by Serde in a Rust app

Question

I have a Rust web on a remote server which I run as:

  RUST_BACKTRACE=1 nohup  /home/my_user/app123 &

When it goes down and I check nohup.log, I see only this:

04:52:22 [WARN] unexpected chunk when body cannot write
04:52:23 [WARN] unexpected chunk when body cannot write
thread 'main' panicked at 'called `Result::unwrap()` on an `Err` value: ErrorImpl { code: EofWhileParsingValue, line: 1, column: 0 }', /checkout/src/libcore/result.rs:906:4
stack backtrace:
   0: <std::time::Instant as core::fmt::Debug>::fmt
   1: <std::time::Instant as core::fmt::Debug>::fmt
   2: <std::time::Instant as core::fmt::Debug>::fmt
   3: <std::time::Instant as core::fmt::Debug>::fmt
   4: <std::time::Instant as core::fmt::Debug>::fmt
   5: <std::time::Instant as core::fmt::Debug>::fmt
   6: <std::time::Instant as core::fmt::Debug>::fmt
   7: <std::time::Instant as core::fmt::Debug>::fmt
   8: <core::cell::BorrowMutError as core::fmt::Debug>::fmt
   9: 
  10: 
  11: 
  12: 
  13: <hyper::version::HttpVersion as core::fmt::Debug>::fmt
  14: 
  15: <std::time::duration::Duration as core::fmt::Debug>::fmt
  16: <std::time::Instant as core::fmt::Debug>::fmt
  17: 
  18: <unknown>

How can I trace the error? It's something related to Serde JSON, but what exactly? There are many places in my app where it could occur.

Why are 9, 10, ... empty? Also without any code, we can't do anything. Can you provide a [mcve](https://stackoverflow.com/help/mcve)? — hellow, Mar 02 '18 at 07:29

score 4 · Accepted Answer · edited Mar 02 '18 at 14:27

There aren't any line numbers in your stack backtrace so the binary running on your server must have been built without debug symbols. As described in How to get a release build with debugging information when using cargo?, add the following section to Cargo.toml of your main binary:

[profile.release]
debug = true

This will include filename and line number information in the binary that can be printed when a panic occurs, making your stack backtraces more useful in the future.

ErrorImpl { code: EofWhileParsingValue, line: 1, column: 0 }

For now, the best we can tell without a backtrace is that you tried to parse an empty JSON string. For example the following triggers the same error.

extern crate serde_json;

fn main() {
    let _: u64 = serde_json::from_str("").unwrap();
}

As a reminder, Result::unwrap() is not an appropriate way to handle errors in a production application. ;-)

`debug = true` --> then why not to use just "cargo build" instead of "cargo build --release"? — jerry, Mar 02 '18 at 10:06
@jerry because then you get no optimizations whatsoever. Not a good idea for a production deployment. — Shepmaster, Mar 02 '18 at 14:18

Unable to trace an error from the backtrace caused by Serde in a Rust app

1 Answers1