Compare commits

...
Author SHA1 Message Date
atlas
2ddabdc410 fixup! fix(#2245): drop TCP fallback from gateway nginx upstream 2026-07-04 22:42:56 +02:00
atlas
32e4c687b8 docs(gateway): remove stale TCP loopback fallback section
The TCP fallback was removed from gateway_nginx.rs in the previous
commit. Update docs/gateway.md to match:
- Step 4: remove 'TCP loopback for agents not yet rebuilt' — always UDS now
- Remove the 'TCP loopback fallback' subsection entirely
2026-07-04 22:42:56 +02:00
atlas
0456206e52 fix(#2245): drop TCP fallback from gateway nginx upstream
Agents run in private netns (always-on isolation). The TCP fallback
to 127.0.0.1:<port> was unreachable from the gateway's host netns
regardless of whether the per-agent socket marker existed.

Remove the conditional entirely: always use the unix socket path.
If the socket is not yet bound, nginx returns 502 which is already
handled by the error_page 502 503 504 = /__hive_agent_unreachable
directive in every location block.

Also removes the unused lifecycle::agent_web_port call and the
now-misleading '.bound state' mention from the render doc comment.
2026-07-04 22:42:56 +02:00
2 changed files with 24 additions and 41 deletions

View file

@ -103,10 +103,10 @@ now set unconditionally for every agent. The mechanism:
(Legacy name `.bound` also accepted during the transition window.)
4. **Gateway side**. `gateway_nginx::write` generates
`/var/lib/hyperhive/gateway/agents.conf` — a plain nginx include
file with one `location /agent/<name>/` block per agent. UDS
upstream (`http://unix:/run/hive-agent/<name>/web.sock:/`) when
`hyperhive-socket-bound` marker present; TCP loopback for agents
that haven't yet been rebuilt under the new config.
file with one `location /agent/<name>/` block per agent. Always
a UDS upstream (`http://unix:/run/hive-agent/<name>/web.sock:/`);
if the socket is not yet bound, nginx returns 502 caught by the
`error_page 502 503 504 = /__hive_agent_unreachable` directive.
The gateway container bind-mounts `/var/lib/hyperhive/gateway/` at
`/run/hive-state/`; nginx includes `/run/hive-state/agents.conf`.
After each write, c0re triggers the appropriate nginx action inside
@ -127,17 +127,6 @@ idempotent — skips the rename when content is unchanged. Failed reloads
are retried automatically on subsequent poll ticks via
`gateway_nginx::reload_if_pending`.
## TCP loopback fallback
While an agent's unix-socket marker is absent, the gateway routes its
`/agent/<name>/` traffic to a TCP loopback upstream in `agents.conf`. The
port is derived on the fly from `lifecycle::agent_web_port(name)` — a pure
FNV-1a hash of the name, reproducible from the name alone, no name
special-cased (so no on-disk port map is needed). Once the agent binds its
unix socket — every agent does, via `HIVE_WEB_SOCKET` — the gateway
switches to the socket upstream from `agent-sockets.json`. The root agent's
UI is routed at `/agent/root/`.
`agents.conf` uses atomic `<path>.tmp` + `rename()` writes so a crashing
c0re process never leaves a partial or unparseable file behind.

View file

@ -1,7 +1,7 @@
//! Runtime nginx include-file generator for the gateway's per-agent
//! `/agent/<name>/` location blocks. Writes
//! `/var/lib/hyperhive/gateway/agents.conf` on every topology change.
//! UDS vs TCP upstream selection, reload trigger (`systemd-run
//! UDS upstream selection, reload trigger (`systemd-run
//! --machine=hive-gateway`), and idempotency:
//! `docs/gateway.md::Per-agent unix-socket upstream`.
@ -14,7 +14,6 @@ use std::time::{SystemTime, UNIX_EPOCH};
use crate::priv_client;
use crate::agent_sockets;
use crate::lifecycle;
/// Set when `write` publishes a new agents.conf; cleared when
/// `reload_gateway_nginx` submits the reload command successfully.
@ -74,12 +73,11 @@ const PROXY_HEADER_BLOCK: &str = " proxy_http_version 1.1;
/// When `frontend_dir` is `None` (legacy) each agent gets a single
/// `location /agent/<name>/` proxy block.
///
/// Output is deterministic for the same (sorted) set of names +
/// `.bound` state: no timestamps, no UUIDs. `BTreeMap` would give
/// alphabetical order; we rely on the caller to pass sorted names if
/// they care about diff stability, but the gateway treats the blocks
/// as unordered by nginx's longest-prefix-match rules so ordering only
/// affects human readability.
/// Output is deterministic for the same (sorted) set of names:
/// no timestamps, no UUIDs. `BTreeMap` would give alphabetical order;
/// we rely on the caller to pass sorted names if they care about diff
/// stability, but the gateway treats the blocks as unordered by nginx's
/// longest-prefix-match rules so ordering only affects human readability.
fn render(names: &[String], frontend_dir: Option<&str>) -> String {
let mut out = String::from(
"# Generated by hive-c0re \u{2014} do not edit.\
@ -87,24 +85,16 @@ fn render(names: &[String], frontend_dir: Option<&str>) -> String {
\n# Reload triggered by hive-c0re via systemd-run --machine=hive-gateway.\n",
);
for name in names {
let port = lifecycle::agent_web_port(name);
// Two upstream forms because named locations (split mode's
// `@<name>_dynamic`) forbid a URI part on `proxy_pass`. The
// legacy prefix-location path keeps the trailing `/` so nginx
// strips `/agent/<name>/` automatically; the named-location
// path strips the prefix via `rewrite` and uses a bare upstream.
let (upstream_prefix, upstream_bare) = if agent_sockets::ready_marker_for(name).exists() {
let sock = agent_sockets::socket_path_for(name).display().to_string();
(
format!("http://unix:{sock}:/"),
format!("http://unix:{sock}:"),
)
} else {
(
format!("http://127.0.0.1:{port}/"),
format!("http://127.0.0.1:{port}"),
)
};
// When the socket is not yet bound, nginx returns 502, caught by
// the error_page directive below.
let sock = agent_sockets::socket_path_for(name).display().to_string();
let upstream_prefix = format!("http://unix:{sock}:/");
let upstream_bare = format!("http://unix:{sock}:");
if let Some(frontend) = frontend_dir {
// Split mode: try to serve files from the nix-store dist first;
@ -303,14 +293,18 @@ mod tests {
}
#[test]
fn render_tcp_upstream_when_no_bound_marker() {
// No .bound file on disk → falls back to TCP loopback.
fn render_uds_upstream_unconditional() {
let names = vec!["iris".to_owned()];
let body = render(&names, None);
let port = lifecycle::agent_web_port("iris");
// UDS form present
assert!(
body.contains(&format!("proxy_pass http://127.0.0.1:{port}/")),
"expected TCP upstream for iris, got:\n{body}"
body.contains("proxy_pass http://unix:"),
"expected UDS upstream for iris, got:\n{body}"
);
// No TCP loopback fallback
assert!(
!body.contains("127.0.0.1"),
"unexpected TCP loopback in output:\n{body}"
);
}