admin: more info in admin GetClusterStatus

doc/drafts/admin-api.md

@@ -69,8 +69,8 @@ Example response body:
 
 ```json
 {
-  "node": "ec79480e0ce52ae26fd00c9da684e4fa56658d9c64cdcecb094e936de0bfe71f",
+  "node": "b10c110e4e854e5aa3f4637681befac755154b20059ec163254ddbfae86b09df",
-  "garageVersion": "git:v0.9.0-dev",
+  "garageVersion": "v0.10.0",
   "garageFeatures": [
     "k2v",
     "sled",
@@ -81,83 +81,92 @@ Example response body:
   ],
   "rustVersion": "1.68.0",
   "dbEngine": "LMDB (using Heed crate)",
-  "knownNodes": [
+  "layoutVersion": 5,
+  "nodes": [
     {
-      "id": "ec79480e0ce52ae26fd00c9da684e4fa56658d9c64cdcecb094e936de0bfe71f",
+      "id": "62b218d848e86a64f7fe1909735f29a4350547b54c4b204f91246a14eb0a1a8c",
-      "addr": "10.0.0.11:3901",
+      "role": {
+        "id": "62b218d848e86a64f7fe1909735f29a4350547b54c4b204f91246a14eb0a1a8c",
+        "zone": "dc1",
+        "capacity": 100000000000,
+        "tags": []
+      },
+      "addr": "10.0.0.3:3901",
+      "hostname": "node3",
       "isUp": true,
-      "lastSeenSecsAgo": 9,
+      "lastSeenSecsAgo": 12,
-      "hostname": "node1"
+      "draining": false,
+      "dataPartition": {
+        "available": 660270088192,
+        "total": 873862266880
+      },
+      "metadataPartition": {
+        "available": 660270088192,
+        "total": 873862266880
+      }
     },
     {
-      "id": "4a6ae5a1d0d33bf895f5bb4f0a418b7dc94c47c0dd2eb108d1158f3c8f60b0ff",
+      "id": "a11c7cf18af297379eff8688360155fe68d9061654449ba0ce239252f5a7487f",
-      "addr": "10.0.0.12:3901",
+      "role": null,
+      "addr": "10.0.0.2:3901",
+      "hostname": "node2",
       "isUp": true,
-      "lastSeenSecsAgo": 1,
+      "lastSeenSecsAgo": 11,
-      "hostname": "node2"
+      "draining": true,
+      "dataPartition": {
+        "available": 660270088192,
+        "total": 873862266880
+      },
+      "metadataPartition": {
+        "available": 660270088192,
+        "total": 873862266880
+      }
     },
     {
-      "id": "23ffd0cdd375ebff573b20cc5cef38996b51c1a7d6dbcf2c6e619876e507cf27",
+      "id": "a235ac7695e0c54d7b403943025f57504d500fdcc5c3e42c71c5212faca040a2",
-      "addr": "10.0.0.21:3901",
+      "role": {
+        "id": "a235ac7695e0c54d7b403943025f57504d500fdcc5c3e42c71c5212faca040a2",
+        "zone": "dc1",
+        "capacity": 100000000000,
+        "tags": []
+      },
+      "addr": "127.0.0.1:3904",
+      "hostname": "lindy",
       "isUp": true,
-      "lastSeenSecsAgo": 7,
+      "lastSeenSecsAgo": 2,
-      "hostname": "node3"
+      "draining": false,
+      "dataPartition": {
+        "available": 660270088192,
+        "total": 873862266880
+      },
+      "metadataPartition": {
+        "available": 660270088192,
+        "total": 873862266880
+      }
     },
     {
-      "id": "e2ee7984ee65b260682086ec70026165903c86e601a4a5a501c1900afe28d84b",
+      "id": "b10c110e4e854e5aa3f4637681befac755154b20059ec163254ddbfae86b09df",
-      "addr": "10.0.0.22:3901",
+      "role": {
+        "id": "b10c110e4e854e5aa3f4637681befac755154b20059ec163254ddbfae86b09df",
+        "zone": "dc1",
+        "capacity": 100000000000,
+        "tags": []
+      },
+      "addr": "10.0.0.1:3901",
+      "hostname": "node1",
       "isUp": true,
-      "lastSeenSecsAgo": 1,
+      "lastSeenSecsAgo": 3,
-      "hostname": "node4"
+      "draining": false,
+      "dataPartition": {
+        "available": 660270088192,
+        "total": 873862266880
+      },
+      "metadataPartition": {
+        "available": 660270088192,
+        "total": 873862266880
+      }
     }
-  ],
+  ]
-  "layout": {
-    "version": 12,
-    "roles": [
-      {
-        "id": "ec79480e0ce52ae26fd00c9da684e4fa56658d9c64cdcecb094e936de0bfe71f",
-        "zone": "dc1",
-        "capacity": 10737418240,
-        "tags": [
-          "node1"
-        ]
-      },
-      {
-        "id": "4a6ae5a1d0d33bf895f5bb4f0a418b7dc94c47c0dd2eb108d1158f3c8f60b0ff",
-        "zone": "dc1",
-        "capacity": 10737418240,
-        "tags": [
-          "node2"
-        ]
-      },
-      {
-        "id": "23ffd0cdd375ebff573b20cc5cef38996b51c1a7d6dbcf2c6e619876e507cf27",
-        "zone": "dc2",
-        "capacity": 10737418240,
-        "tags": [
-          "node3"
-        ]
-      }
-    ],
-    "stagedRoleChanges": [
-      {
-        "id": "e2ee7984ee65b260682086ec70026165903c86e601a4a5a501c1900afe28d84b",
-        "remove": false,
-        "zone": "dc2",
-        "capacity": 10737418240,
-        "tags": [
-          "node4"
-        ]
-      }
-      {
-        "id": "23ffd0cdd375ebff573b20cc5cef38996b51c1a7d6dbcf2c6e619876e507cf27",
-        "remove": true,
-        "zone": null,
-        "capacity": null,
-        "tags": null,
-      }
-    ]
-  }
 }
 ```
 

src/api/admin/cluster.rs

@@ -1,3 +1,4 @@
+use std::collections::HashMap;
 use std::net::SocketAddr;
 use std::sync::Arc;
 
@@ -15,25 +16,95 @@ use crate::admin::error::*;
 use crate::helpers::{json_ok_response, parse_json_body};
 
 pub async fn handle_get_cluster_status(garage: &Arc<Garage>) -> Result<Response<Body>, Error> {
+	let layout = garage.system.cluster_layout();
+	let mut nodes = garage
+		.system
+		.get_known_nodes()
+		.into_iter()
+		.map(|i| {
+			(
+				i.id,
+				NodeResp {
+					id: hex::encode(i.id),
+					addr: Some(i.addr),
+					hostname: i.status.hostname,
+					is_up: i.is_up,
+					last_seen_secs_ago: i.last_seen_secs_ago,
+					data_partition: i
+						.status
+						.data_disk_avail
+						.map(|(avail, total)| FreeSpaceResp {
+							available: avail,
+							total,
+						}),
+					metadata_partition: i.status.meta_disk_avail.map(|(avail, total)| {
+						FreeSpaceResp {
+							available: avail,
+							total,
+						}
+					}),
+					..Default::default()
+				},
+			)
+		})
+		.collect::<HashMap<_, _>>();
+
+	for (id, _, role) in layout.current().roles.items().iter() {
+		if let layout::NodeRoleV(Some(r)) = role {
+			let role = NodeRoleResp {
+				id: hex::encode(id),
+				zone: r.zone.to_string(),
+				capacity: r.capacity,
+				tags: r.tags.clone(),
+			};
+			match nodes.get_mut(id) {
+				None => {
+					nodes.insert(
+						*id,
+						NodeResp {
+							id: hex::encode(id),
+							role: Some(role),
+							..Default::default()
+						},
+					);
+				}
+				Some(n) => {
+					if n.role.is_none() {
+						n.role = Some(role);
+					}
+				}
+			}
+		}
+	}
+
+	for ver in layout.versions.iter().rev().skip(1) {
+		for (id, _, role) in ver.roles.items().iter() {
+			if let layout::NodeRoleV(Some(r)) = role {
+				if !nodes.contains_key(id) && r.capacity.is_some() {
+					nodes.insert(
+						*id,
+						NodeResp {
+							id: hex::encode(id),
+							draining: true,
+							..Default::default()
+						},
+					);
+				}
+			}
+		}
+	}
+
+	let mut nodes = nodes.into_iter().map(|(_, v)| v).collect::<Vec<_>>();
+	nodes.sort_by(|x, y| x.id.cmp(&y.id));
+
 	let res = GetClusterStatusResponse {
 		node: hex::encode(garage.system.id),
 		garage_version: garage_util::version::garage_version(),
 		garage_features: garage_util::version::garage_features(),
 		rust_version: garage_util::version::rust_version(),
 		db_engine: garage.db.engine(),
-		known_nodes: garage
+		layout_version: layout.current().version,
-			.system
+		nodes,
-			.get_known_nodes()
-			.into_iter()
-			.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: format_cluster_layout(&garage.system.cluster_layout()),
 	};
 
 	Ok(json_ok_response(&res)?)
@@ -157,8 +228,8 @@ struct GetClusterStatusResponse {
 	garage_features: Option<&'static [&'static str]>,
 	rust_version: &'static str,
 	db_engine: String,
-	known_nodes: Vec<KnownNodeResp>,
+	layout_version: u64,
-	layout: GetClusterLayoutResponse,
+	nodes: Vec<NodeResp>,
 }
 
 #[derive(Serialize)]
@@ -192,14 +263,27 @@ struct NodeRoleResp {
 	tags: Vec<String>,
 }
 
-#[derive(Serialize)]
+#[derive(Serialize, Default)]
 #[serde(rename_all = "camelCase")]
-struct KnownNodeResp {
+struct FreeSpaceResp {
+	available: u64,
+	total: u64,
+}
+
+#[derive(Serialize, Default)]
+#[serde(rename_all = "camelCase")]
+struct NodeResp {
 	id: String,
-	addr: SocketAddr,
+	role: Option<NodeRoleResp>,
+	addr: Option<SocketAddr>,
+	hostname: Option<String>,
 	is_up: bool,
 	last_seen_secs_ago: Option<u64>,
-	hostname: String,
+	draining: bool,
+	#[serde(skip_serializing_if = "Option::is_none")]
+	data_partition: Option<FreeSpaceResp>,
+	#[serde(skip_serializing_if = "Option::is_none")]
+	metadata_partition: Option<FreeSpaceResp>,
 }
 
 // ---- update functions ----

src/garage/admin/mod.rs

@@ -295,7 +295,7 @@ impl AdminRpcHandler {
 			let info = node_info.get(id);
 			let status = info.map(|x| &x.status);
 			let role = layout.current().roles.get(id).and_then(|x| x.0.as_ref());
-			let hostname = status.map(|x| x.hostname.as_str()).unwrap_or("?");
+			let hostname = status.and_then(|x| x.hostname.as_deref()).unwrap_or("?");
 			let zone = role.map(|x| x.zone.as_str()).unwrap_or("?");
 			let capacity = role
 				.map(|x| x.capacity_string())

src/garage/cli/cmd.rs

@@ -62,6 +62,7 @@ pub async fn cmd_status(rpc_cli: &Endpoint<SystemRpc, ()>, rpc_host: NodeID) ->
 	let mut healthy_nodes =
 		vec!["ID\tHostname\tAddress\tTags\tZone\tCapacity\tDataAvail".to_string()];
 	for adv in status.iter().filter(|adv| adv.is_up) {
+		let host = adv.status.hostname.as_deref().unwrap_or("?");
 		if let Some(NodeRoleV(Some(cfg))) = layout.current().roles.get(&adv.id) {
 			let data_avail = match &adv.status.data_disk_avail {
 				_ if cfg.capacity.is_none() => "N/A".into(),
@@ -75,7 +76,7 @@ pub async fn cmd_status(rpc_cli: &Endpoint<SystemRpc, ()>, rpc_host: NodeID) ->
 			healthy_nodes.push(format!(
 				"{id:?}\t{host}\t{addr}\t[{tags}]\t{zone}\t{capacity}\t{data_avail}",
 				id = adv.id,
-				host = adv.status.hostname,
+				host = host,
 				addr = adv.addr,
 				tags = cfg.tags.join(","),
 				zone = cfg.zone,
@@ -95,7 +96,7 @@ pub async fn cmd_status(rpc_cli: &Endpoint<SystemRpc, ()>, rpc_host: NodeID) ->
 				healthy_nodes.push(format!(
 					"{id:?}\t{host}\t{addr}\t[{tags}]\t{zone}\tdraining metadata...",
 					id = adv.id,
-					host = adv.status.hostname,
+					host = host,
 					addr = adv.addr,
 					tags = cfg.tags.join(","),
 					zone = cfg.zone,
@@ -108,7 +109,7 @@ pub async fn cmd_status(rpc_cli: &Endpoint<SystemRpc, ()>, rpc_host: NodeID) ->
 				healthy_nodes.push(format!(
 					"{id:?}\t{h}\t{addr}\t\t\t{new_role}",
 					id = adv.id,
-					h = adv.status.hostname,
+					h = host,
 					addr = adv.addr,
 					new_role = new_role,
 				));
@@ -149,7 +150,7 @@ pub async fn cmd_status(rpc_cli: &Endpoint<SystemRpc, ()>, rpc_host: NodeID) ->
 			// it is in a failed state, add proper line to the output
 			let (host, addr, last_seen) = match adv {
 				Some(adv) => (
-					adv.status.hostname.as_str(),
+					adv.status.hostname.as_deref().unwrap_or("?"),
 					adv.addr.to_string(),
 					adv.last_seen_secs_ago
 						.map(|s| tf.convert(Duration::from_secs(s)))

src/rpc/system.rs

@@ -126,7 +126,7 @@ pub struct System {
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct NodeStatus {
 	/// Hostname of the node
-	pub hostname: String,
+	pub hostname: Option<String>,
 
 	/// Replication factor configured on the node
 	pub replication_factor: usize,
@@ -765,9 +765,11 @@ impl EndpointHandler<SystemRpc> for System {
 impl NodeStatus {
 	fn initial(replication_factor: usize, layout_manager: &LayoutManager) -> Self {
 		NodeStatus {
-			hostname: gethostname::gethostname()
+			hostname: Some(
-				.into_string()
+				gethostname::gethostname()
-				.unwrap_or_else(|_| "<invalid utf-8>".to_string()),
+					.into_string()
+					.unwrap_or_else(|_| "<invalid utf-8>".to_string()),
+			),
 			replication_factor,
 			layout_digest: layout_manager.layout().digest(),
 			meta_disk_avail: None,
@@ -777,7 +779,7 @@ impl NodeStatus {
 
 	fn unknown() -> Self {
 		NodeStatus {
-			hostname: "?".to_string(),
+			hostname: None,
 			replication_factor: 0,
 			layout_digest: Default::default(),
 			meta_disk_avail: None,