Merge pull request 'disable_scrub configuration option' (#774) from disable-scrub into main

Reviewed-on: Deuxfleurs/garage#774

doc/book/operations/durability-repairs.md

@@ -19,7 +19,7 @@ connecting to. To run on all nodes, add the `-a` flag as follows:
 
 # Data block operations
 
-## Data store scrub
+## Data store scrub {#scrub}
 
 Scrubbing the data store means examining each individual data block to check that
 their content is correct, by verifying their hash. Any block found to be corrupted

doc/book/reference-manual/configuration.md

@@ -14,6 +14,7 @@ metadata_dir = "/var/lib/garage/meta"
 data_dir = "/var/lib/garage/data"
 metadata_fsync = true
 data_fsync = false
+disable_scrub = false
 
 db_engine = "lmdb"
 
@@ -87,6 +88,7 @@ Top-level configuration options:
 [`data_dir`](#data_dir),
 [`data_fsync`](#data_fsync),
 [`db_engine`](#db_engine),
+[`disable_scrub`](#disable_scrub),
 [`lmdb_map_size`](#lmdb_map_size),
 [`metadata_dir`](#metadata_dir),
 [`metadata_fsync`](#metadata_fsync),
@@ -344,6 +346,22 @@ at the cost of a moderate drop in write performance.
 Similarly to `metatada_fsync`, this is likely not necessary
 if geographical replication is used.
 
+#### `disable_scrub` {#disable_scrub}
+
+By default, Garage runs a scrub of the data directory approximately once per
+month, with a random delay to avoid all nodes running at the same time.  When
+it scrubs the data directory, Garage will read all of the data files stored on
+disk to check their integrity, and will rebuild any data files that it finds
+corrupted, using the remaining valid copies stored on other nodes.
+See [this page](@/documentation/operations/durability-repair.md#scrub) for details.
+
+Set the `disable_scrub` configuration value to `true` if you don't need Garage
+to scrub the data directory, for instance if you are already scrubbing at the
+filesystem level. Note that in this case, if you find a corrupted data file,
+you should delete it from the data directory and then call `garage repair
+blocks` on the node to ensure that it re-obtains a copy from another node on
+the network.
+
 #### `block_size` {#block_size}
 
 Garage splits stored objects in consecutive chunks of size `block_size`

src/block/manager.rs

@@ -22,7 +22,7 @@ use garage_net::stream::{read_stream_to_end, stream_asyncread, ByteStream};
 use garage_db as db;
 
 use garage_util::background::{vars, BackgroundRunner};
-use garage_util::config::DataDirEnum;
+use garage_util::config::Config;
 use garage_util::data::*;
 use garage_util::error::*;
 use garage_util::metrics::RecordDuration;
@@ -84,6 +84,7 @@ pub struct BlockManager {
 
 	data_fsync: bool,
 	compression_level: Option<i32>,
+	disable_scrub: bool,
 
 	mutation_lock: Vec<Mutex<BlockManagerLocked>>,
 
@@ -119,9 +120,7 @@ struct BlockManagerLocked();
 impl BlockManager {
 	pub fn new(
 		db: &db::Db,
-		data_dir: DataDirEnum,
+		config: &Config,
-		data_fsync: bool,
-		compression_level: Option<i32>,
 		replication: TableShardedReplication,
 		system: Arc<System>,
 	) -> Result<Arc<Self>, Error> {
@@ -131,11 +130,13 @@ impl BlockManager {
 		let data_layout = match data_layout_persister.load() {
 			Ok(mut layout) => {
 				layout
-					.update(&data_dir)
+					.update(&config.data_dir)
 					.ok_or_message("invalid data_dir config")?;
 				layout
 			}
-			Err(_) => DataLayout::initialize(&data_dir).ok_or_message("invalid data_dir config")?,
+			Err(_) => {
+				DataLayout::initialize(&config.data_dir).ok_or_message("invalid data_dir config")?
+			}
 		};
 		data_layout_persister
 			.save(&data_layout)
@@ -154,7 +155,7 @@ impl BlockManager {
 			.endpoint("garage_block/manager.rs/Rpc".to_string());
 
 		let metrics = BlockManagerMetrics::new(
-			compression_level,
+			config.compression_level,
 			rc.rc.clone(),
 			resync.queue.clone(),
 			resync.errors.clone(),
@@ -166,8 +167,9 @@ impl BlockManager {
 			replication,
 			data_layout: ArcSwap::new(Arc::new(data_layout)),
 			data_layout_persister,
-			data_fsync,
+			data_fsync: config.data_fsync,
-			compression_level,
+			disable_scrub: config.disable_scrub,
+			compression_level: config.compression_level,
 			mutation_lock: vec![(); MUTEX_COUNT]
 				.iter()
 				.map(|_| Mutex::new(BlockManagerLocked()))
@@ -194,6 +196,7 @@ impl BlockManager {
 		}
 
 		// Spawn scrub worker
+		if !self.disable_scrub {
 			let (scrub_tx, scrub_rx) = mpsc::channel(1);
 			self.tx_scrub_command.store(Some(Arc::new(scrub_tx)));
 			bg.spawn_worker(ScrubWorker::new(
@@ -202,10 +205,12 @@ impl BlockManager {
 				self.scrub_persister.clone(),
 			));
 		}
+	}
 
 	pub fn register_bg_vars(&self, vars: &mut vars::BgVars) {
 		self.resync.register_bg_vars(vars);
 
+		if !self.disable_scrub {
 			vars.register_rw(
 				&self.scrub_persister,
 				"scrub-tranquility",
@@ -222,6 +227,7 @@ impl BlockManager {
 				p.get_with(|x| x.corruptions_detected)
 			});
 		}
+	}
 
 	/// Ask nodes that might have a (possibly compressed) block for it
 	/// Return it as a stream with a header

src/model/garage.rs

@@ -177,14 +177,7 @@ impl Garage {
 		};
 
 		info!("Initialize block manager...");
-		let block_manager = BlockManager::new(
+		let block_manager = BlockManager::new(&db, &config, data_rep_param, system.clone())?;
-			&db,
-			config.data_dir.clone(),
-			config.data_fsync,
-			config.compression_level,
-			data_rep_param,
-			system.clone(),
-		)?;
 		block_manager.register_bg_vars(&mut bg_vars);
 
 		// ---- admin tables ----

src/util/config.rs

@@ -23,6 +23,10 @@ pub struct Config {
 	#[serde(default)]
 	pub data_fsync: bool,
 
+	/// Disable automatic scrubbing of the data directory
+	#[serde(default)]
+	pub disable_scrub: bool,
+
 	/// Size of data blocks to save to disk
 	#[serde(
 		deserialize_with = "deserialize_capacity",