lfg2025/archy
1
0
Fork 0
Code Issues 11 Pull Requests Actions Packages Projects Releases 1 Wiki Activity

fix(indeehub): preserve reviewed runtimes across reconciliation and lifecycle

Browse Source
This commit is contained in:
archipelago
2026-10-07 18:51:32 -04:00
parent d19124f5dc
commit b0b95810e0
5 changed files with 312 additions and 15 deletions
Download Patch File Download Diff File Expand all files Collapse all files
core/archipelago/src/api/rpc/package/runtime.rs
+29 -12
View File
@@ -111,7 +111,7 @@ impl RpcHandler {
let _lifecycle_guard = lifecycle_guard;
let _op_guard = op_lock.lock().await;
let result = if let Some(orchestrator) = orchestrator.as_ref() {
do_orchestrator_package_start(orchestrator.as_ref(), &to_start).await
do_orchestrator_package_start(orchestrator.as_ref(), &to_start, &data_dir).await
} else {
do_package_start(&to_start).await
};
@@ -239,11 +239,12 @@ impl RpcHandler {
.await;
let op_lock = app_op_lock(package_id);
let data_dir = self.config.data_dir.clone();
tokio::spawn(async move {
let _lifecycle_guard = lifecycle_guard;
let _op_guard = op_lock.lock().await;
let result = if let Some(orchestrator) = orchestrator.as_ref() {
do_orchestrator_package_stop(orchestrator.as_ref(), &to_stop).await
do_orchestrator_package_stop(orchestrator.as_ref(), &to_stop, &data_dir).await
} else {
do_package_stop(&containers).await
};
@@ -348,7 +349,7 @@ impl RpcHandler {
let _lifecycle_guard = lifecycle_guard;
let _op_guard = op_lock.lock().await;
let result = if let Some(orchestrator) = orchestrator.as_ref() {
do_orchestrator_package_restart(orchestrator.as_ref(), &to_restart).await
do_orchestrator_package_restart(orchestrator.as_ref(), &to_restart, &data_dir).await
} else {
do_package_restart(&containers).await
};
@@ -872,17 +873,24 @@ async fn do_package_start(to_start: &[String]) -> Result<()> {
async fn do_orchestrator_package_start(
orchestrator: &dyn crate::container::traits::ContainerOrchestrator,
to_start: &[String],
data_dir: &std::path::Path,
) -> Result<()> {
for name in to_start {
orchestrator.validate_start(name).await?;
}
let mut errors = Vec::new();
for (i, name) in to_start.iter().enumerate() {
if i > 0 {
tokio::time::sleep(std::time::Duration::from_secs(2)).await;
}
repair_before_package_start(name).await;
wait_before_package_start(name).await;
let managed = crate::container::supervised_update::installed_unit(data_dir, name)?.is_some();
if !managed {
repair_before_package_start(name).await;
wait_before_package_start(name).await;
}
match orchestrator.start(name).await {
Ok(()) => wait_after_orchestrator_start(name).await,
Err(e) if is_unknown_app_id_error(&e) => {
Err(e) if !managed && is_unknown_app_id_error(&e) => {
// Collect instead of `?`: aborting here skipped every later
// stack member (mempool gate flakes #73/#74 — see
// order_present_containers).
@@ -902,7 +910,9 @@ async fn do_orchestrator_package_start(
Err(anyhow::anyhow!("Start failed: {}", errors.join("; ")))
} else {
for name in to_start {
ensure_runtime_host_port_listener(name).await?;
if crate::container::supervised_update::installed_unit(data_dir, name)?.is_none() {
ensure_runtime_host_port_listener(name).await?;
}
}
Ok(())
}
@@ -1131,12 +1141,14 @@ async fn podman_start_container(container_name: &str) -> Result<Output> {
async fn do_orchestrator_package_stop(
orchestrator: &dyn crate::container::traits::ContainerOrchestrator,
containers: &[String],
data_dir: &std::path::Path,
) -> Result<()> {
let mut errors = Vec::new();
for name in containers {
let managed = crate::container::supervised_update::installed_unit(data_dir, name)?.is_some();
match orchestrator.stop(name).await {
Ok(()) => {}
Err(e) if is_unknown_app_id_error(&e) => {
Err(e) if !managed && is_unknown_app_id_error(&e) => {
if let Err(e) = do_package_stop(&[name.clone()]).await {
errors.push(format!("{}: {:#}", name, e));
}
@@ -1145,7 +1157,7 @@ async fn do_orchestrator_package_stop(
// stop removes the --rm container before the orchestrator's
// defensive `podman stop` fallback probes it, and a stack member
// can be absent outright (stopped earlier, or never revived).
Err(e) if is_missing_container_error(&format!("{:#}", e)) => {
Err(e) if !managed && is_missing_container_error(&format!("{:#}", e)) => {
tracing::debug!(container = %name, error = %e, "stop: container already absent — treating as stopped");
}
Err(e) => {
@@ -1231,7 +1243,7 @@ async fn cascade_restart_address_caching_dependents(
let prev = flip_package_state(state_manager, dep, PackageState::Restarting).await;
let target = vec![dep.to_string()];
let result = if let Some(orch) = orchestrator {
do_orchestrator_package_restart(orch.as_ref(), &target).await
do_orchestrator_package_restart(orch.as_ref(), &target, data_dir).await
} else {
do_package_restart(&target).await
};
@@ -1299,9 +1311,14 @@ fn uses_single_orchestrator_app(package_id: &str) -> bool {
async fn do_orchestrator_package_restart(
orchestrator: &dyn crate::container::traits::ContainerOrchestrator,
to_restart: &[String],
data_dir: &std::path::Path,
) -> Result<()> {
do_orchestrator_package_stop(orchestrator, to_restart).await?;
do_orchestrator_package_start(orchestrator, to_restart).await
for name in to_restart {
orchestrator.validate_start(name).await?;
}
let stop_order: Vec<String> = to_restart.iter().rev().cloned().collect();
do_orchestrator_package_stop(orchestrator, &stop_order, data_dir).await?;
do_orchestrator_package_start(orchestrator, to_restart, data_dir).await
}
/// Stop all containers with their per-container graceful-shutdown timeout.