Retry on tar extraction errors (#9753)
Some checks are pending
CI / check system | alpine (push) Blocked by required conditions
CI / check system | python on macos aarch64 (push) Blocked by required conditions
CI / check system | homebrew python on macos aarch64 (push) Blocked by required conditions
CI / check system | python on macos x86_64 (push) Blocked by required conditions
CI / check system | python3.10 on windows (push) Blocked by required conditions
CI / Determine changes (push) Waiting to run
CI / lint (push) Waiting to run
CI / cargo clippy | ubuntu (push) Blocked by required conditions
CI / cargo clippy | windows (push) Blocked by required conditions
CI / cargo dev generate-all (push) Blocked by required conditions
CI / cargo shear (push) Waiting to run
CI / check windows trampoline | i686 (push) Blocked by required conditions
CI / cargo test | ubuntu (push) Blocked by required conditions
CI / cargo test | macos (push) Blocked by required conditions
CI / cargo test | windows (push) Blocked by required conditions
CI / check windows trampoline | aarch64 (push) Blocked by required conditions
CI / check windows trampoline | x86_64 (push) Blocked by required conditions
CI / test windows trampoline | i686 (push) Blocked by required conditions
CI / test windows trampoline | x86_64 (push) Blocked by required conditions
CI / typos (push) Waiting to run
CI / mkdocs (push) Waiting to run
CI / build binary | linux (push) Blocked by required conditions
CI / build binary | macos aarch64 (push) Blocked by required conditions
CI / build binary | macos x86_64 (push) Blocked by required conditions
CI / build binary | windows (push) Blocked by required conditions
CI / cargo build (msrv) (push) Blocked by required conditions
CI / build binary | freebsd (push) Blocked by required conditions
CI / ecosystem test | prefecthq/prefect (push) Blocked by required conditions
CI / ecosystem test | pallets/flask (push) Blocked by required conditions
CI / integration test | conda on ubuntu (push) Blocked by required conditions
CI / integration test | free-threaded on linux (push) Blocked by required conditions
CI / integration test | free-threaded on windows (push) Blocked by required conditions
CI / integration test | pypy on ubuntu (push) Blocked by required conditions
CI / integration test | pypy on windows (push) Blocked by required conditions
CI / integration test | graalpy on ubuntu (push) Blocked by required conditions
CI / integration test | graalpy on windows (push) Blocked by required conditions
CI / integration test | github actions (push) Blocked by required conditions
CI / integration test | determine publish changes (push) Blocked by required conditions
CI / integration test | uv publish (push) Blocked by required conditions
CI / check cache | ubuntu (push) Blocked by required conditions
CI / check cache | macos aarch64 (push) Blocked by required conditions
CI / check system | python on debian (push) Blocked by required conditions
CI / check system | python on fedora (push) Blocked by required conditions
CI / check system | python on ubuntu (push) Blocked by required conditions
CI / check system | python on opensuse (push) Blocked by required conditions
CI / check system | python on rocky linux 8 (push) Blocked by required conditions
CI / check system | python on rocky linux 9 (push) Blocked by required conditions
CI / check system | pypy on ubuntu (push) Blocked by required conditions
CI / check system | pyston (push) Blocked by required conditions
CI / check system | python3.10 on windows x86 (push) Blocked by required conditions
CI / check system | python3.13 on windows (push) Blocked by required conditions
CI / check system | python3.12 via chocolatey (push) Blocked by required conditions
CI / check system | python3.9 via pyenv (push) Blocked by required conditions
CI / check system | python3.13 (push) Blocked by required conditions
CI / check system | conda3.11 on linux (push) Blocked by required conditions
CI / check system | conda3.8 on linux (push) Blocked by required conditions
CI / check system | conda3.11 on macos (push) Blocked by required conditions
CI / check system | conda3.8 on macos (push) Blocked by required conditions
CI / check system | conda3.11 on windows (push) Blocked by required conditions
CI / check system | conda3.8 on windows (push) Blocked by required conditions
CI / check system | amazonlinux (push) Blocked by required conditions
CI / check system | embedded python3.10 on windows (push) Blocked by required conditions
CI / benchmarks (push) Blocked by required conditions

## Summary

So the error here is:

```rust
ExtractError("cpython-3.11.11%2B20241206-aarch64-apple-darwin-install_only_stripped.tar.gz", Io(Custom { kind: UnexpectedEof, error: TarError { desc: "failed to unpack `/Users/crmarsh/.local/share/uv/python/.cache/.tmpkqFzqE/python/lib/libpython3.11.dylib`", io: Custom { kind: UnexpectedEof, error: TarError { desc: "failed to unpack `python/lib/libpython3.11.dylib` into `/Users/crmarsh/.local/share/uv/python/.cache/.tmpkqFzqE/python/lib/libpython3.11.dylib`", io: Custom { kind: UnexpectedEof, error: "unexpected end of file" } } } } }))
```

This isn't a Reqwest error, so we miss it in
`is_extended_transient_error`.

We could add `TarError` or `ExtractError` here, but... should we? This
PR just extends it to any error that has an IO source. I don't see much
of a downside.

Closes https://github.com/astral-sh/uv/issues/9747.

## Test Plan

First, ran: `uv run ./scripts/create-python-mirror.py --name cpython
--arch aarch64 --os darwin`.

Then, dropped this into `./scripts/mirror/server.py`:

```python
import os
import random
from http.server import SimpleHTTPRequestHandler, HTTPServer


class GlitchyStaticServer(SimpleHTTPRequestHandler):
    def do_GET(self):
        """Handle GET request."""
        file_path = self.translate_path(self.path)
        
        if not os.path.exists(file_path):
            self.send_error(404, "File not found")
            return
        
        try:
            with open(file_path, 'rb') as f:
                file_content = f.read()

            # Introduce an "unexpected end of file" glitch randomly
            if random.random() < 0.75:  # 75% chance of glitch
                glitch_point = random.randint(1, len(file_content) - 1)
                file_content = file_content[:glitch_point]

            self.send_response(200)
            self.send_header("Content-type", self.guess_type(file_path))
            self.send_header("Content-Length", len(file_content))
            self.end_headers()
            self.wfile.write(file_content)
        
        except Exception as e:
            self.send_error(500, f"Internal Server Error: {e}")
        

def run(server_class=HTTPServer, handler_class=GlitchyStaticServer, port=8080):
    """Run the server."""
    server_address = ('', port)
    httpd = server_class(server_address, handler_class)
    print(f"Serving on port {port} with glitchy behavior")
    httpd.serve_forever()


if __name__ == "__main__":
    run()
```

Then ran `python server.py` from that directory.

From there, ran `UV_PYTHON_INSTALL_MIRROR="http://localhost:8080" cargo
run python install 3.11 --reinstall --verbose` to reliably test retries.
This commit is contained in:
Charlie Marsh 2024-12-10 07:33:08 -05:00 committed by GitHub
parent 85a4fb4471
commit eb21e4bd25
No known key found for this signature in database
GPG key ID: B5690EEEBB952194

View file

@ -25,7 +25,7 @@ use uv_warnings::warn_user_once;
use crate::linehaul::LineHaul;
use crate::middleware::OfflineMiddleware;
use crate::tls::read_identity;
use crate::{Connectivity, WrappedReqwestError};
use crate::Connectivity;
pub const DEFAULT_RETRIES: u32 = 3;
@ -458,51 +458,16 @@ impl RetryableStrategy for UvRetryableStrategy {
pub fn is_extended_transient_error(err: &dyn Error) -> bool {
trace!("Attempting to retry error: {err:?}");
if let Some(err) = find_source::<WrappedReqwestError>(&err) {
// First, look for `WrappedReqwestError`, which wraps `reqwest::Error` but doesn't always
// include it in the source.
if let Some(io) = find_source::<std::io::Error>(&err) {
if io.kind() == std::io::ErrorKind::ConnectionReset
|| io.kind() == std::io::ErrorKind::UnexpectedEof
{
trace!(
"Retrying error: `ConnectionReset` or `UnexpectedEof` (`WrappedReqwestError`)"
);
return true;
}
trace!("Cannot retry error: not one of `ConnectionReset` or `UnexpectedEof` (`WrappedReqwestError`)");
} else {
trace!("Cannot retry error: not an IO error (`WrappedReqwestError`)");
}
} else if let Some(err) = find_source::<reqwest_middleware::Error>(&err) {
// Next, look for `reqwest_middleware::Error`, which wraps `reqwest::Error`, but also
// includes errors from the middleware stack.
if let Some(io) = find_source::<std::io::Error>(&err) {
if io.kind() == std::io::ErrorKind::ConnectionReset
|| io.kind() == std::io::ErrorKind::UnexpectedEof
{
trace!("Retrying error: `ConnectionReset` or `UnexpectedEof` (`reqwest_middleware::Error`)");
return true;
}
trace!("Cannot retry error: not one of `ConnectionReset` or `UnexpectedEof` (`reqwest_middleware::Error`)");
} else {
trace!("Cannot retry error: not an IO error (`reqwest_middleware::Error`)");
}
} else if let Some(err) = find_source::<reqwest::Error>(&err) {
// Finally, look for `reqwest::Error`, which is the most common error type.
if let Some(io) = find_source::<std::io::Error>(&err) {
if io.kind() == std::io::ErrorKind::ConnectionReset
|| io.kind() == std::io::ErrorKind::UnexpectedEof
{
trace!("Retrying error: `ConnectionReset` or `UnexpectedEof` (`reqwest::Error`)");
return true;
}
trace!("Cannot retry error: not one of `ConnectionReset` or `UnexpectedEof` (`reqwest::Error`)");
} else {
trace!("Cannot retry error: not an IO error (`reqwest::Error`)");
if let Some(io) = find_source::<std::io::Error>(&err) {
if io.kind() == std::io::ErrorKind::ConnectionReset
|| io.kind() == std::io::ErrorKind::UnexpectedEof
{
trace!("Retrying error: `ConnectionReset` or `UnexpectedEof`");
return true;
}
trace!("Cannot retry error: not one of `ConnectionReset` or `UnexpectedEof`");
} else {
trace!("Cannot retry error: not a reqwest error");
trace!("Cannot retry error: not an IO error");
}
false