chore: release v1.7.61-alpha

core/Cargo.lock

@@ -80,7 +80,7 @@ checksum = "a23eb6b1614318a8071c9b2521f36b424b2c83db5eb3a0fead4a6c0809af6e61"
 
 [[package]]
 name = "archipelago"
-version = "1.7.60-alpha"
+version = "1.7.61-alpha"
 dependencies = [
  "anyhow",
  "archipelago-container",

core/archipelago/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "archipelago"
-version = "1.7.60-alpha"
+version = "1.7.61-alpha"
 edition = "2021"
 description = "Archipelago Bitcoin Node OS - Native backend"
 authors = ["Archipelago Team"]

core/archipelago/src/server.rs

@@ -840,6 +840,20 @@ const CONTAINER_ABSENCE_THRESHOLD: u32 = 3;
 /// scanner's authoritative view. Applies to all transitional variants.
 const TRANSITIONAL_STUCK_TIMEOUT: Duration = Duration::from_secs(120);
 
+/// Multi-container installs can legitimately spend several minutes before the
+/// primary user-facing container exists. BTCPay, for example, pulls/starts
+/// Postgres and NBXplorer before `btcpay-server`; do not erase its installing
+/// card just because the primary container is absent during that setup window.
+const INSTALLING_STUCK_TIMEOUT: Duration = Duration::from_secs(20 * 60);
+
+fn transitional_stuck_timeout(state: &crate::data_model::PackageState) -> Duration {
+    if *state == crate::data_model::PackageState::Installing {
+        INSTALLING_STUCK_TIMEOUT
+    } else {
+        TRANSITIONAL_STUCK_TIMEOUT
+    }
+}
+
 /// Returns true if `state` is one of the transitional variants that a
 /// `spawn_transitional`-style background task owns. While such a state is
 /// set, the package scanner must not overwrite it with whatever podman
@@ -961,13 +975,14 @@ async fn scan_and_update_packages(
         let overwrite = match existing {
             Some(existing_entry) if is_transitional(&existing_entry.state) => {
                 let entered = *transitional_since.entry(id.clone()).or_insert(now);
-                let stuck = now.duration_since(entered) > TRANSITIONAL_STUCK_TIMEOUT;
+                let timeout = transitional_stuck_timeout(&existing_entry.state);
+                let stuck = now.duration_since(entered) > timeout;
                 if stuck {
                     warn!(
                         "Container {} stuck in {:?} for >{}s; overriding with scan state {:?}",
                         id,
                         existing_entry.state,
-                        TRANSITIONAL_STUCK_TIMEOUT.as_secs(),
+                        timeout.as_secs(),
                         pkg.state
                     );
                     transitional_since.remove(id);
@@ -1015,12 +1030,13 @@ async fn scan_and_update_packages(
             if let Some(entry) = merged.get(&id) {
                 if is_transitional(&entry.state) {
                     let entered = *transitional_since.entry(id.clone()).or_insert(now);
-                    if now.duration_since(entered) > TRANSITIONAL_STUCK_TIMEOUT {
+                    let timeout = transitional_stuck_timeout(&entry.state);
+                    if now.duration_since(entered) > timeout {
                         warn!(
                             "Container {} stuck in {:?} and absent for >{}s; removing stale transitional state",
                             id,
                             entry.state,
-                            TRANSITIONAL_STUCK_TIMEOUT.as_secs()
+                            timeout.as_secs()
                         );
                         merged.remove(&id);
                         transitional_since.remove(&id);
@@ -1247,4 +1263,13 @@ mod merge_tests {
             assert!(!is_transitional(&s), "{:?} should NOT be transitional", s);
         }
     }
+
+    #[test]
+    fn installing_uses_longer_stale_timeout_than_other_transitions() {
+        assert!(transitional_stuck_timeout(&PackageState::Installing) > TRANSITIONAL_STUCK_TIMEOUT);
+        assert_eq!(
+            transitional_stuck_timeout(&PackageState::Stopping),
+            TRANSITIONAL_STUCK_TIMEOUT
+        );
+    }
 }