Compare commits

...

2 Commits

Author SHA1 Message Date
archipelago
36015a19fe docs(master-plan): §8b session-b state — connection-lost+netbird+UX-merge shipped to .228, uninstall ghost fix, workstream F in progress
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-23 15:26:17 -04:00
archipelago
e57514b690 fix(uninstall): never ghost a removed app in My Apps on cleanup residue
handle_package_uninstall lumped every teardown failure into one `errors` vec
and returned Err on any of them BEFORE removing the package state entry — so a
non-fatal cleanup hiccup (a slow/failed `sudo rm -rf` of a large data dir, a
volume/network removal) left the app's containers gone but its entry in
package_data → a ghost in My Apps, and the spawned task reverted it to Installed.

Split the failures: container removal that even force-rm can't complete (app
genuinely still present) keeps the entry + returns Err; everything after the
containers are gone is best-effort. Remove the state entry as soon as the
containers are gone — BEFORE the slow volume/data teardown — so My Apps updates
immediately and residue can never ghost the app. set_uninstall_stage is a no-op
once the entry is gone (if-let guard), so the later stages don't re-create it.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-06-23 15:23:16 -04:00
2 changed files with 95 additions and 41 deletions

View File

@ -312,7 +312,16 @@ impl RpcHandler {
let mut stopped = 0u32;
let mut removed = 0u32;
let mut errors = Vec::new();
// Two distinct failure classes, kept separate so they don't get
// conflated (the old single `errors` vec did, which caused the "ghost in
// My Apps" bug): `container_errors` means a container could NOT be
// removed (force-rm failed too) — the app is genuinely still present, so
// we keep its state entry and surface a hard error. `cleanup_errors`
// means volume/network/data-dir teardown left residue — the containers
// are already gone, so the app IS uninstalled and MUST disappear from My
// Apps; the residue is logged but never ghosts the app.
let mut container_errors: Vec<String> = Vec::new();
let mut cleanup_errors: Vec<String> = Vec::new();
self.set_uninstall_stage(
package_id,
@ -370,7 +379,7 @@ impl RpcHandler {
let msg =
format!("Failed to remove {}: {}; {}", name, stderr.trim(), e);
tracing::error!("Uninstall {}: {}", package_id, msg);
errors.push(msg);
container_errors.push(msg);
}
}
}
@ -379,12 +388,35 @@ impl RpcHandler {
Err(force_err) => {
let msg = format!("Failed to remove {}: {}; {}", name, e, force_err);
tracing::error!("Uninstall {}: {}", package_id, msg);
errors.push(msg);
container_errors.push(msg);
}
},
}
}
// A container that survived even force-remove means the app is NOT
// actually uninstalled — keep its state entry and fail so the spawned
// task reverts it to its prior state (and the user can retry), rather
// than orphaning a live container that's missing from My Apps.
if !container_errors.is_empty() {
tracing::error!(
"Uninstall {}: containers could not be removed: {:?}",
package_id,
container_errors
);
return Err(anyhow::anyhow!(
"Uninstall {} failed: {}",
package_id,
container_errors.join("; ")
));
}
// Containers are gone → the app is uninstalled. Remove its state entry
// NOW, before the (possibly slow, possibly fallible) volume/data
// teardown below, so My Apps updates immediately and a residue failure
// can never leave a ghost. Reinstall/scan no longer see a stale entry.
self.remove_package_state_entry(package_id).await;
self.set_uninstall_stage(package_id, "Cleaning up volumes")
.await;
// Avoid global Podman volume prune on production nodes: store-wide
@ -432,45 +464,55 @@ impl RpcHandler {
let stderr = String::from_utf8_lossy(&o.stderr);
let msg = format!("Failed to remove data {}: {}", dir, stderr.trim());
tracing::error!("Uninstall {}: {}", package_id, msg);
errors.push(msg);
cleanup_errors.push(msg);
}
Err(e) => {
let msg = format!("Failed to remove data {}: {}", dir, e);
tracing::error!("Uninstall {}: {}", package_id, msg);
errors.push(msg);
cleanup_errors.push(msg);
}
_ => {}
}
}
}
if !errors.is_empty() {
// The app is already gone from My Apps (entry removed above). Residual
// volume/data cleanup failures are logged but NEVER ghost the app — a
// reinstall and the next uninstall both tolerate leftover dirs.
if !cleanup_errors.is_empty() {
tracing::error!(
"Uninstall {} completed with errors: {:?}",
"Uninstall {} removed but left cleanup residue: {:?}",
package_id,
errors
cleanup_errors
);
return Err(anyhow::anyhow!(
"Uninstall {} partially failed: {}",
package_id,
errors.join("; ")
));
}
tracing::info!(
"Uninstall {} complete: stopped={}, removed={}",
"Uninstall {} complete: stopped={}, removed={}, cleanup_errors={}",
package_id,
stopped,
removed
removed,
cleanup_errors.len()
);
// Immediately remove from in-memory state so the UI updates without
// waiting for the scanner's absence threshold (3 scans × 60s each).
{
Ok(serde_json::json!({
"status": "uninstalled",
"stopped": stopped,
"removed": removed,
"cleanup_warnings": cleanup_errors,
}))
}
/// Remove a package's entry (and any alias keys) from persisted state so it
/// disappears from My Apps immediately, without waiting for the scanner's
/// absence threshold (3 scans × 60s). Called as soon as an uninstall has
/// removed the app's containers — before the slower volume/data teardown —
/// so a residue failure can never leave a ghost entry behind.
async fn remove_package_state_entry(&self, package_id: &str) {
let (mut data, _rev) = self.state_manager.get_snapshot().await;
let before = data.package_data.len();
data.package_data.remove(package_id);
// Also remove any alias keys (e.g. "bitcoin-knots" vs "bitcoin")
// Also remove any alias keys (e.g. "bitcoin-knots" vs "bitcoin").
let aliases: Vec<String> = data
.package_data
.keys()
@ -489,13 +531,6 @@ impl RpcHandler {
}
}
Ok(serde_json::json!({
"status": "uninstalled",
"stopped": stopped,
"removed": removed,
}))
}
/// Start a bundled app (create container from pre-loaded image if needed).
pub(in crate::api::rpc) async fn handle_bundled_app_start(
&self,

View File

@ -245,7 +245,26 @@ phases 26 (`dual-ecash-design.md`).
## 8b. SESSION STATE + RESUME (updated 2026-06-23) — READ §8b "CURRENT STATE + RESUME" FIRST
### ▶ CURRENT STATE + RESUME (2026-06-23) — RESUME FROM HERE (works from any device)
### ▶ SESSION b (2026-06-23 PM) — LATEST, RESUME FROM HERE
**Canonical resume detail: memory `project_session_resume_2026_06_23b` (▶️ top of MEMORY.md).**
`gitea-vps2/main = 4346007d` pushed; local HEAD `e57514b6` (uninstall fix, committed, **not pushed/deployed**).
Shipped + verified live on .228 (all in 4346007d):
- **Connection-lost FULLY fixed** — companion `image_exists` journal-flood (Stdio::null) + netbird UDP-port reconcile churn (`wait_for_manifest_host_ports` tcp-only). .228: flood→0, ws/db→0 disconnects, load 3.95→2.26.
- **netbird → manifest-driven** (#20 ph4) — 3 manifests + 4 orchestrator primitives (base64 secret, GeneratedCert+`ensure_manifest_certs`, templated-file render `{{HOST_IP}}/{{NETWORK_GATEWAY}}/{{secret:}}`, udp port protocol). Live: https 8087→200, OIDC→200, resolver=gateway. Legacy-Rust delete deferred to post-full-verify.
- **registry-manifest flip (code)**`EMBED_MANIFESTS` default-on, `main.rs` bounded pre-load `refresh_catalog`. Catalog regenerated w/ 52 embedded manifests but **NOT published** (gitignored + never committed; publish = force-add to gitea-vps2 main). Do after fleet binary roll.
- **UX regression root-caused + fixed** — the mobile/desktop UX (loader/AppLoadingScreen, store-driven launch, app icons, android webview footer) was on `companion-mobile-ux` and **never merged to main**, so any main build silently dropped it. **Merged → main**, frontend redeployed to .228. Android 0.4.9/code13 pushed for user to build APK elsewhere.
In progress — **Workstream F lifecycle bugs** (this §, user-picked next):
- **uninstall ghost — FIXED (e57514b6), not deployed.** `handle_package_uninstall` returned Err on any cleanup-residue failure *before* removing the package state entry → ghost in My Apps + revert-to-Installed. Now: split container vs cleanup errors; remove state entry as soon as containers gone (before slow data rm). NEXT: deploy + verify a real uninstall clears My Apps (CAUTION: destroys app data — get user OK).
- grafana reinstall-stops (#14, likely same root cause, re-test after) + fedimint guardian wait-vs-stuck (#15, not started).
Next: deploy+verify uninstall fix on .228 → push e57514b6 → roll binary to fleet → publish catalog → finish F → Phase 3 → multinode.
---
### ▶ CURRENT STATE + RESUME (2026-06-23) — earlier session-a baseline (historical)
**✅ HEADLINE (2026-06-23): single-node gate GREEN (`run-gate.sh` 5/5 on .228, 0 not-ok) +
multinode test deploy DONE to 6 nodes.** The exit criterion (§5) is met. Green took fixing **two real