admin api: change cluster status/layout to use lists and not maps (fix #377)

3 files changed, 175 insertions, 75 deletions

diff --git a/doc/drafts/admin-api.md b/doc/drafts/admin-api.md
index e0252f71..b1a8f402 100644
--- a/doc/drafts/admin-api.md
+++ b/doc/drafts/admin-api.md
@@ -56,7 +56,7 @@ See `/v0/health` for an API that also returns JSON output.
 
 ### Cluster operations
 
-#### GetClusterStatus `GET /v0/status`
+#### GetClusterStatus `GET /v1/status`
 
 Returns the cluster's current status in JSON, including:
 
@@ -70,67 +70,93 @@ Example response body:
 ```json
 {
   "node": "ec79480e0ce52ae26fd00c9da684e4fa56658d9c64cdcecb094e936de0bfe71f",
-  "garage_version": "git:v0.8.0",
-  "knownNodes": {
-    "ec79480e0ce52ae26fd00c9da684e4fa56658d9c64cdcecb094e936de0bfe71f": {
+  "garageVersion": "git:v0.9.0-dev",
+  "garageFeatures": [
+    "k2v",
+    "sled",
+    "lmdb",
+    "sqlite",
+    "metrics",
+    "bundled-libs"
+  ],
+  "rustVersion": "1.68.0",
+  "dbEngine": "LMDB (using Heed crate)",
+  "knownNodes": [
+    {
+      "id": "ec79480e0ce52ae26fd00c9da684e4fa56658d9c64cdcecb094e936de0bfe71f",
       "addr": "10.0.0.11:3901",
       "is_up": true,
       "last_seen_secs_ago": 9,
       "hostname": "node1"
     },
-    "4a6ae5a1d0d33bf895f5bb4f0a418b7dc94c47c0dd2eb108d1158f3c8f60b0ff": {
+    {
+      "id": "4a6ae5a1d0d33bf895f5bb4f0a418b7dc94c47c0dd2eb108d1158f3c8f60b0ff",
       "addr": "10.0.0.12:3901",
       "is_up": true,
       "last_seen_secs_ago": 1,
       "hostname": "node2"
     },
-    "23ffd0cdd375ebff573b20cc5cef38996b51c1a7d6dbcf2c6e619876e507cf27": {
+    {
+      "id": "23ffd0cdd375ebff573b20cc5cef38996b51c1a7d6dbcf2c6e619876e507cf27",
       "addr": "10.0.0.21:3901",
       "is_up": true,
       "last_seen_secs_ago": 7,
       "hostname": "node3"
     },
-    "e2ee7984ee65b260682086ec70026165903c86e601a4a5a501c1900afe28d84b": {
+    {
+      "id": "e2ee7984ee65b260682086ec70026165903c86e601a4a5a501c1900afe28d84b",
       "addr": "10.0.0.22:3901",
       "is_up": true,
       "last_seen_secs_ago": 1,
       "hostname": "node4"
     }
-  },
+  ],
   "layout": {
     "version": 12,
-    "roles": {
-      "ec79480e0ce52ae26fd00c9da684e4fa56658d9c64cdcecb094e936de0bfe71f": {
+    "roles": [
+      {
+        "id": "ec79480e0ce52ae26fd00c9da684e4fa56658d9c64cdcecb094e936de0bfe71f",
         "zone": "dc1",
-        "capacity": 4,
+        "capacity": 10737418240,
         "tags": [
           "node1"
         ]
       },
-      "4a6ae5a1d0d33bf895f5bb4f0a418b7dc94c47c0dd2eb108d1158f3c8f60b0ff": {
+      {
+        "id": "4a6ae5a1d0d33bf895f5bb4f0a418b7dc94c47c0dd2eb108d1158f3c8f60b0ff",
         "zone": "dc1",
-        "capacity": 6,
+        "capacity": 10737418240,
         "tags": [
           "node2"
         ]
       },
-      "23ffd0cdd375ebff573b20cc5cef38996b51c1a7d6dbcf2c6e619876e507cf27": {
+      {
+        "id": "23ffd0cdd375ebff573b20cc5cef38996b51c1a7d6dbcf2c6e619876e507cf27",
         "zone": "dc2",
-        "capacity": 10,
+        "capacity": 10737418240,
         "tags": [
           "node3"
         ]
       }
-    },
-    "stagedRoleChanges": {
-      "e2ee7984ee65b260682086ec70026165903c86e601a4a5a501c1900afe28d84b": {
+    ],
+    "stagedRoleChanges": [
+      {
+        "id": "e2ee7984ee65b260682086ec70026165903c86e601a4a5a501c1900afe28d84b",
+        "remove": false,
         "zone": "dc2",
-        "capacity": 5,
+        "capacity": 10737418240,
         "tags": [
           "node4"
         ]
       }
-    }
+      {
+        "id": "23ffd0cdd375ebff573b20cc5cef38996b51c1a7d6dbcf2c6e619876e507cf27",
+        "remove": true,
+        "zone": null,
+        "capacity": null,
+        "tags": null,
+      }
+    ]
   }
 }
 ```
@@ -198,7 +224,7 @@ Example response:
 ]
 ```
 
-#### GetClusterLayout `GET /v0/layout`
+#### GetClusterLayout `GET /v1/layout`
 
 Returns the cluster's current layout in JSON, including:
 
@@ -212,42 +238,54 @@ Example response body:
 ```json
 {
   "version": 12,
-  "roles": {
-    "ec79480e0ce52ae26fd00c9da684e4fa56658d9c64cdcecb094e936de0bfe71f": {
+  "roles": [
+    {
+      "id": "ec79480e0ce52ae26fd00c9da684e4fa56658d9c64cdcecb094e936de0bfe71f",
       "zone": "dc1",
-      "capacity": 4,
+      "capacity": 10737418240,
       "tags": [
         "node1"
       ]
     },
-    "4a6ae5a1d0d33bf895f5bb4f0a418b7dc94c47c0dd2eb108d1158f3c8f60b0ff": {
+    {
+      "id": "4a6ae5a1d0d33bf895f5bb4f0a418b7dc94c47c0dd2eb108d1158f3c8f60b0ff",
       "zone": "dc1",
-      "capacity": 6,
+      "capacity": 10737418240,
       "tags": [
         "node2"
       ]
     },
-    "23ffd0cdd375ebff573b20cc5cef38996b51c1a7d6dbcf2c6e619876e507cf27": {
+    {
+      "id": "23ffd0cdd375ebff573b20cc5cef38996b51c1a7d6dbcf2c6e619876e507cf27",
       "zone": "dc2",
-      "capacity": 10,
+      "capacity": 10737418240,
       "tags": [
         "node3"
       ]
     }
-  },
-  "stagedRoleChanges": {
-    "e2ee7984ee65b260682086ec70026165903c86e601a4a5a501c1900afe28d84b": {
+  ],
+  "stagedRoleChanges": [
+    {
+      "id": "e2ee7984ee65b260682086ec70026165903c86e601a4a5a501c1900afe28d84b",
+      "remove": false,
       "zone": "dc2",
-      "capacity": 5,
+      "capacity": 10737418240,
       "tags": [
         "node4"
       ]
     }
-  }
+    {
+      "id": "23ffd0cdd375ebff573b20cc5cef38996b51c1a7d6dbcf2c6e619876e507cf27",
+      "remove": true,
+      "zone": null,
+      "capacity": null,
+      "tags": null,
+    }
+  ]
 }
 ```
 
-#### UpdateClusterLayout `POST /v0/layout`
+#### UpdateClusterLayout `POST /v1/layout`
 
 Send modifications to the cluster layout. These modifications will
 be included in the staged role changes, visible in subsequent calls
@@ -259,8 +297,9 @@ the layout.
 Request body format:
 
 ```json
-{
-  <node_id>: {
+[
+  {
+    "id": <node_id>,
     "capacity": <new_capacity>,
     "zone": <new_zone>,
     "tags": [
@@ -268,9 +307,11 @@ Request body format:
       ...
     ]
   },
-  <node_id_to_remove>: null,
-  ...
-}
+  {
+    "id": <node_id_to_remove>,
+    "remove": true
+  }
+]
 ```
 
 Contrary to the CLI that may update only a subset of the fields
diff --git a/src/api/admin/cluster.rs b/src/api/admin/cluster.rs
index a2c97ee5..8a208a2c 100644
--- a/src/api/admin/cluster.rs
+++ b/src/api/admin/cluster.rs
@@ -1,4 +1,3 @@
-use std::collections::HashMap;
 use std::net::SocketAddr;
 use std::sync::Arc;
 
@@ -8,7 +7,7 @@ use serde::{Deserialize, Serialize};
 use garage_util::crdt::*;
 use garage_util::data::*;
 
-use garage_rpc::layout::*;
+use garage_rpc::layout;
 
 use garage_model::garage::Garage;
 
@@ -26,16 +25,12 @@ pub async fn handle_get_cluster_status(garage: &Arc<Garage>) -> Result<Response<
 			.system
 			.get_known_nodes()
 			.into_iter()
-			.map(|i| {
-				(
-					hex::encode(i.id),
-					KnownNodeResp {
-						addr: i.addr,
-						is_up: i.is_up,
-						last_seen_secs_ago: i.last_seen_secs_ago,
-						hostname: i.status.hostname,
-					},
-				)
+			.map(|i| KnownNodeResp {
+				id: hex::encode(i.id),
+				addr: i.addr,
+				is_up: i.is_up,
+				last_seen_secs_ago: i.last_seen_secs_ago,
+				hostname: i.status.hostname,
 			})
 			.collect(),
 		layout: get_cluster_layout(garage),
@@ -82,25 +77,49 @@ pub async fn handle_get_cluster_layout(garage: &Arc<Garage>) -> Result<Response<
 fn get_cluster_layout(garage: &Arc<Garage>) -> GetClusterLayoutResponse {
 	let layout = garage.system.get_cluster_layout();
 
+	let roles = layout
+		.roles
+		.items()
+		.iter()
+		.filter_map(|(k, _, v)| v.0.clone().map(|x| (k, x)))
+		.map(|(k, v)| NodeRoleResp {
+			id: hex::encode(k),
+			zone: v.zone.clone(),
+			capacity: v.capacity,
+			tags: v.tags.clone(),
+		})
+		.collect::<Vec<_>>();
+
+	let staged_role_changes = layout
+		.staging_roles
+		.items()
+		.iter()
+		.filter(|(k, _, v)| layout.roles.get(k) != Some(v))
+		.map(|(k, _, v)| match &v.0 {
+			None => NodeRoleChange {
+				id: hex::encode(k),
+				remove: true,
+				..Default::default()
+			},
+			Some(r) => NodeRoleChange {
+				id: hex::encode(k),
+				remove: false,
+				zone: Some(r.zone.clone()),
+				capacity: r.capacity,
+				tags: Some(r.tags.clone()),
+			},
+		})
+		.collect::<Vec<_>>();
+
 	GetClusterLayoutResponse {
 		version: layout.version,
-		roles: layout
-			.roles
-			.items()
-			.iter()
-			.filter(|(_, _, v)| v.0.is_some())
-			.map(|(k, _, v)| (hex::encode(k), v.0.clone()))
-			.collect(),
-		staged_role_changes: layout
-			.staging_roles
-			.items()
-			.iter()
-			.filter(|(k, _, v)| layout.roles.get(k) != Some(v))
-			.map(|(k, _, v)| (hex::encode(k), v.0.clone()))
-			.collect(),
+		roles,
+		staged_role_changes,
 	}
 }
 
+// ----
+
 #[derive(Serialize)]
 #[serde(rename_all = "camelCase")]
 struct GetClusterStatusResponse {
@@ -109,7 +128,7 @@ struct GetClusterStatusResponse {
 	garage_features: Option<&'static [&'static str]>,
 	rust_version: &'static str,
 	db_engine: String,
-	known_nodes: HashMap<String, KnownNodeResp>,
+	known_nodes: Vec<KnownNodeResp>,
 	layout: GetClusterLayoutResponse,
 }
 
@@ -124,19 +143,31 @@ struct ConnectClusterNodesResponse {
 #[serde(rename_all = "camelCase")]
 struct GetClusterLayoutResponse {
 	version: u64,
-	roles: HashMap<String, Option<NodeRole>>,
-	staged_role_changes: HashMap<String, Option<NodeRole>>,
+	roles: Vec<NodeRoleResp>,
+	staged_role_changes: Vec<NodeRoleChange>,
+}
+
+#[derive(Serialize)]
+#[serde(rename_all = "camelCase")]
+struct NodeRoleResp {
+	id: String,
+	zone: String,
+	capacity: Option<u64>,
+	tags: Vec<String>,
 }
 
 #[derive(Serialize)]
 #[serde(rename_all = "camelCase")]
 struct KnownNodeResp {
+	id: String,
 	addr: SocketAddr,
 	is_up: bool,
 	last_seen_secs_ago: Option<u64>,
 	hostname: String,
 }
 
+// ---- update functions ----
+
 pub async fn handle_update_cluster_layout(
 	garage: &Arc<Garage>,
 	req: Request<Body>,
@@ -148,13 +179,23 @@ pub async fn handle_update_cluster_layout(
 	let mut roles = layout.roles.clone();
 	roles.merge(&layout.staging_roles);
 
-	for (node, role) in updates {
-		let node = hex::decode(node).ok_or_bad_request("Invalid node identifier")?;
+	for change in updates {
+		let node = hex::decode(&change.id).ok_or_bad_request("Invalid node identifier")?;
 		let node = Uuid::try_from(&node).ok_or_bad_request("Invalid node identifier")?;
 
+		let new_role = match (change.remove, change.zone, change.capacity, change.tags) {
+			(true, None, None, None) => None,
+			(false, Some(zone), capacity, Some(tags)) => Some(layout::NodeRole {
+				zone,
+				capacity,
+				tags,
+			}),
+			_ => return Err(Error::bad_request("Invalid layout change")),
+		};
+
 		layout
 			.staging_roles
-			.merge(&roles.update_mutator(node, NodeRoleV(role)));
+			.merge(&roles.update_mutator(node, layout::NodeRoleV(new_role)));
 	}
 
 	garage.system.update_cluster_layout(&layout).await?;
@@ -196,10 +237,28 @@ pub async fn handle_revert_cluster_layout(
 		.body(Body::empty())?)
 }
 
-type UpdateClusterLayoutRequest = HashMap<String, Option<NodeRole>>;
+// ----
+
+type UpdateClusterLayoutRequest = Vec<NodeRoleChange>;
 
 #[derive(Deserialize)]
 #[serde(rename_all = "camelCase")]
 struct ApplyRevertLayoutRequest {
 	version: u64,
 }
+
+// ----
+
+#[derive(Serialize, Deserialize, Default)]
+#[serde(rename_all = "camelCase")]
+struct NodeRoleChange {
+	id: String,
+	#[serde(default)]
+	remove: bool,
+	#[serde(default)]
+	zone: Option<String>,
+	#[serde(default)]
+	capacity: Option<u64>,
+	#[serde(default)]
+	tags: Option<Vec<String>>,
+}
diff --git a/src/api/admin/router.rs b/src/api/admin/router.rs
index 0dcb1546..5af3ffb5 100644
--- a/src/api/admin/router.rs
+++ b/src/api/admin/router.rs
@@ -95,12 +95,12 @@ impl Endpoint {
 			GET "/check" => CheckWebsiteEnabled,
 			GET "/health" => Health,
 			GET "/metrics" => Metrics,
-			GET "/v0/status" => GetClusterStatus,
+			GET "/v1/status" => GetClusterStatus,
 			GET "/v0/health" => GetClusterHealth,
 			POST "/v0/connect" => ConnectClusterNodes,
 			// Layout endpoints
-			GET "/v0/layout" => GetClusterLayout,
-			POST "/v0/layout" => UpdateClusterLayout,
+			GET "/v1/layout" => GetClusterLayout,
+			POST "/v1/layout" => UpdateClusterLayout,
 			POST "/v0/layout/apply" => ApplyClusterLayout,
 			POST "/v0/layout/revert" => RevertClusterLayout,
 			// API key endpoints