FlockError Unknown error 524

This is on NERSC Cori compute nodes (which I suspect @davidanthoff may have access to, so mentioning this). The error message is:

```
Juliaup configuration is locked by another process, waiting for it to unlock.
```
with crash report
```
name = 'Juliaup launcher'
operating_system = 'unix:Unknown'
crate_version = '1.5.36'
explanation = '''
Panic occurred in file 'src/config_file.rs' at line 108
'''
cause = 'called `Result::unwrap()` on an `Err` value: FlockError { data: File { fd: 3, path: "/global/u1/m/marius/.julia-cori/juliaup/.juliaup-lock", read: true, write: true }, err: Os { code: 524, kind: Uncategorized, message: "Unknown error 524" } }'
method = 'Panic'
backtrace = '''

0: 0x555555593413 - julialauncher::run_app::h01f348944e7e5962
1: 0x55555559a839 - julialauncher::main::h23a700439e8fb0e7
2: 0x55555558ca93 - std::sys_common::backtrace::__rust_begin_short_backtrace::h130f4c622a35468f
3: 0x55555559d6f7 - main
4: 0x2aaaab53a34a - __libc_start_main
5: 0x55555558c969 - _start
6: 0x0 - '''
```

I checked and I can both touch and rm the lock file, but this error persists even after removing. This does not happen on login nodes of this cluster.

Maybe not super helpful, but I also found this which I suspect is the same thing happening to an HDF5 library https://github.com/ALPSCore/ALPSCore/issues/348 maybe there's something helpful in there. It must have to do with how the home directory is mounted on these compute nodes, but in any case, it would be nice if juliaup still worked there somehow.

Start at src/config_file.rs line 108 and the julialauncher entry point, then reproduce the lock failure on a NERSC Cori compute node and compare it with a login node. Done means juliaup handles the mounted-home locking error without panicking and still works on the compute node.

Written by the indexing model from the issue text.