Merge branch 'main' into next-0.10

Cargo.lock

@@ -1350,6 +1350,7 @@ dependencies = [
  "sha2",
  "static_init",
  "structopt",
+ "syslog-tracing",
  "timeago",
  "tokio",
  "toml",
@@ -3985,6 +3986,17 @@ dependencies = [
  "unicode-xid",
 ]
 
+[[package]]
+name = "syslog-tracing"
+version = "0.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "340b1540dcdb6b066bc2966e7974f977ab1a38f21b2be189014ffb0cc2405768"
+dependencies = [
+ "libc",
+ "tracing-core",
+ "tracing-subscriber",
+]
+
 [[package]]
 name = "system-configuration"
 version = "0.5.1"

Cargo.nix

@@ -34,7 +34,7 @@ args@{
   ignoreLockHash,
 }:
 let
-  nixifiedLockHash = "a49da9d5ef560672a34c1e004c0122e706a74fac512300f20858f136cd00582e";
+  nixifiedLockHash = "1ef5e578c148e63bdc6491d497aba66b38dcf011779d417228906ce7b19d55f4";
   workspaceSrc = if args.workspaceSrc == null then ./. else args.workspaceSrc;
   currentLockHash = builtins.hashFile "sha256" (workspaceSrc + /Cargo.lock);
   lockHashIgnored = if ignoreLockHash
@@ -1927,6 +1927,8 @@ in
       (lib.optional (rootFeatures' ? "garage/default" || rootFeatures' ? "garage/metrics" || rootFeatures' ? "garage/opentelemetry-prometheus") "opentelemetry-prometheus")
       (lib.optional (rootFeatures' ? "garage/default" || rootFeatures' ? "garage/metrics" || rootFeatures' ? "garage/prometheus") "prometheus")
       (lib.optional (rootFeatures' ? "garage/default" || rootFeatures' ? "garage/sqlite") "sqlite")
+      (lib.optional (rootFeatures' ? "garage/syslog") "syslog")
+      (lib.optional (rootFeatures' ? "garage/syslog" || rootFeatures' ? "garage/syslog-tracing") "syslog-tracing")
       (lib.optional (rootFeatures' ? "garage/system-libs") "system-libs")
       (lib.optional (rootFeatures' ? "garage/telemetry-otlp") "telemetry-otlp")
     ];
@@ -1960,6 +1962,7 @@ in
       serde_bytes = (rustPackages."registry+https://github.com/rust-lang/crates.io-index".serde_bytes."0.11.14" { inherit profileName; }).out;
       sha1 = (rustPackages."registry+https://github.com/rust-lang/crates.io-index".sha1."0.10.6" { inherit profileName; }).out;
       structopt = (rustPackages."registry+https://github.com/rust-lang/crates.io-index".structopt."0.3.26" { inherit profileName; }).out;
+      ${ if rootFeatures' ? "garage/syslog" || rootFeatures' ? "garage/syslog-tracing" then "syslog_tracing" else null } = (rustPackages."registry+https://github.com/rust-lang/crates.io-index".syslog-tracing."0.3.0" { inherit profileName; }).out;
       timeago = (rustPackages."registry+https://github.com/rust-lang/crates.io-index".timeago."0.4.2" { inherit profileName; }).out;
       tokio = (rustPackages."registry+https://github.com/rust-lang/crates.io-index".tokio."1.36.0" { inherit profileName; }).out;
       toml = (rustPackages."registry+https://github.com/rust-lang/crates.io-index".toml."0.8.10" { inherit profileName; }).out;
@@ -5675,6 +5678,18 @@ in
     };
   });
   
+  "registry+https://github.com/rust-lang/crates.io-index".syslog-tracing."0.3.0" = overridableMkRustCrate (profileName: rec {
+    name = "syslog-tracing";
+    version = "0.3.0";
+    registry = "registry+https://github.com/rust-lang/crates.io-index";
+    src = fetchCratesIo { inherit name version; sha256 = "340b1540dcdb6b066bc2966e7974f977ab1a38f21b2be189014ffb0cc2405768"; };
+    dependencies = {
+      ${ if rootFeatures' ? "garage/syslog" || rootFeatures' ? "garage/syslog-tracing" then "libc" else null } = (rustPackages."registry+https://github.com/rust-lang/crates.io-index".libc."0.2.153" { inherit profileName; }).out;
+      ${ if rootFeatures' ? "garage/syslog" || rootFeatures' ? "garage/syslog-tracing" then "tracing_core" else null } = (rustPackages."registry+https://github.com/rust-lang/crates.io-index".tracing-core."0.1.32" { inherit profileName; }).out;
+      ${ if rootFeatures' ? "garage/syslog" || rootFeatures' ? "garage/syslog-tracing" then "tracing_subscriber" else null } = (rustPackages."registry+https://github.com/rust-lang/crates.io-index".tracing-subscriber."0.3.18" { inherit profileName; }).out;
+    };
+  });
+  
   "registry+https://github.com/rust-lang/crates.io-index".system-configuration."0.5.1" = overridableMkRustCrate (profileName: rec {
     name = "system-configuration";
     version = "0.5.1";

Cargo.toml

@@ -76,6 +76,7 @@ kuska-handshake = { version = "0.2.0", features = ["default", "async_std"] }
 clap = { version = "4.1", features = ["derive", "env"] }
 pretty_env_logger = "0.5"
 structopt = { version = "0.3", default-features = false }
+syslog-tracing = "0.3"
 tracing = "0.1"
 tracing-subscriber = { version = "0.3", features = ["env-filter"] }
 

doc/book/cookbook/from-source.md

@@ -90,5 +90,6 @@ The following feature flags are available in v0.8.0:
 | `kubernetes-discovery` | optional | Enable automatic registration and discovery<br>of cluster nodes through the Kubernetes API |
 | `metrics` | *by default* | Enable collection of metrics in Prometheus format on the admin API |
 | `telemetry-otlp` | optional | Enable collection of execution traces using OpenTelemetry |
+| `syslog` | optional | Enable logging to Syslog |
 | `lmdb` | *by default* | Enable using LMDB to store Garage's metadata |
 | `sqlite` | *by default* | Enable using Sqlite3 to store Garage's metadata |

doc/book/reference-manual/configuration.md

@@ -31,6 +31,8 @@ rpc_bind_addr = "[::]:3901"
 rpc_bind_outgoing = false
 rpc_public_addr = "[fc00:1::1]:3901"
 
+allow_world_readable_secrets = false
+
 bootstrap_peers = [
     "563e1ac825ee3323aa441e72c26d1030d6d4414aeb3dd25287c531e7fc2bc95d@[fc00:1::1]:3901",
     "86f0f26ae4afbd59aaf9cfb059eefac844951efd5b8caeec0d53f4ed6c85f332@[fc00:1::2]:3901",
@@ -81,7 +83,10 @@ The following gives details about each available configuration option.
 
 ### Index
 
+[Environment variables](#env_variables).
+
 Top-level configuration options:
+[`allow_world_readable_secrets`](#allow_world_readable_secrets),
 [`block_size`](#block_size),
 [`bootstrap_peers`](#bootstrap_peers),
 [`compression_level`](#compression_level),
@@ -132,6 +137,23 @@ The `[admin]` section:
 [`admin_token`/`admin_token_file`](#admin_token),
 [`trace_sink`](#admin_trace_sink),
 
+### Environment variables {#env_variables}
+
+The following configuration parameter must be specified as an environment
+variable, it does not exist in the configuration file:
+
+- `GARAGE_LOG_TO_SYSLOG` (since v0.9.4): set this to `1` or `true` to make the
+  Garage daemon send its logs to `syslog` (using the libc `syslog` function)
+  instead of printing to stderr.
+
+The following environment variables can be used to override the corresponding
+values in the configuration file:
+
+- [`GARAGE_ALLOW_WORLD_READABLE_SECRETS`](#allow_world_readable_secrets)
+- [`GARAGE_RPC_SECRET` and `GARAGE_RPC_SECRET_FILE`](#rpc_secret)
+- [`GARAGE_ADMIN_TOKEN` and `GARAGE_ADMIN_TOKEN_FILE`](#admin_token)
+- [`GARAGE_METRICS_TOKEN` and `GARAGE_METRICS_TOKEN`](#admin_metrics_token)
+
 
 ### Top-level configuration options
 
@@ -502,7 +524,7 @@ be obtained by running `garage node id` and then included directly in the
 key will be returned by `garage node id` and you will have to add the IP
 yourself.
 
-### `allow_world_readable_secrets`
+### `allow_world_readable_secrets` or `GARAGE_ALLOW_WORLD_READABLE_SECRETS` (env) {#allow_world_readable_secrets}
 
 Garage checks the permissions of your secret files to make sure they're not
 world-readable. In some cases, the check might fail and consider your files as

k2v_test.py

@@ -1,158 +0,0 @@
-#!/usr/bin/env python
-
-import os
-import requests
-from datetime import datetime
-
-# let's talk to our AWS Elasticsearch cluster
-#from requests_aws4auth import AWS4Auth
-#auth = AWS4Auth('GK31c2f218a2e44f485b94239e',
-#                       'b892c0665f0ada8a4755dae98baa3b133590e11dae3bcc1f9d769d67f16c3835',
-#                       'us-east-1',
-#                       's3')
-
-from aws_requests_auth.aws_auth import AWSRequestsAuth
-auth = AWSRequestsAuth(aws_access_key='GK31c2f218a2e44f485b94239e',
-        aws_secret_access_key='b892c0665f0ada8a4755dae98baa3b133590e11dae3bcc1f9d769d67f16c3835',
-        aws_host='localhost:3812',
-        aws_region='us-east-1',
-        aws_service='k2v')
-
-
-print("-- ReadIndex")
-response = requests.get('http://localhost:3812/alex',
-                        auth=auth)
-print(response.headers)
-print(response.text)
-
-
-sort_keys = ["a", "b", "c", "d"]
-
-for sk in sort_keys:
-    print("-- (%s) Put initial (no CT)"%sk)
-    response = requests.put('http://localhost:3812/alex/root?sort_key=%s'%sk,
-                            auth=auth,
-                            data='{}: Hello, world!'.format(datetime.timestamp(datetime.now())))
-    print(response.headers)
-    print(response.text)
-
-    print("-- Get")
-    response = requests.get('http://localhost:3812/alex/root?sort_key=%s'%sk,
-                            auth=auth)
-    print(response.headers)
-    print(response.text)
-    ct = response.headers["x-garage-causality-token"]
-
-    print("-- ReadIndex")
-    response = requests.get('http://localhost:3812/alex',
-                            auth=auth)
-    print(response.headers)
-    print(response.text)
-
-    print("-- Put with CT")
-    response = requests.put('http://localhost:3812/alex/root?sort_key=%s'%sk,
-                            auth=auth,
-                            headers={'x-garage-causality-token': ct},
-                            data='{}: Good bye, world!'.format(datetime.timestamp(datetime.now())))
-    print(response.headers)
-    print(response.text)
-
-    print("-- Get")
-    response = requests.get('http://localhost:3812/alex/root?sort_key=%s'%sk,
-                            auth=auth)
-    print(response.headers)
-    print(response.text)
-
-    print("-- Put again with same CT (concurrent)")
-    response = requests.put('http://localhost:3812/alex/root?sort_key=%s'%sk,
-                            auth=auth,
-                            headers={'x-garage-causality-token': ct},
-                            data='{}: Concurrent value, oops'.format(datetime.timestamp(datetime.now())))
-    print(response.headers)
-    print(response.text)
-
-for sk in sort_keys:
-    print("-- (%s) Get"%sk)
-    response = requests.get('http://localhost:3812/alex/root?sort_key=%s'%sk,
-                            auth=auth)
-    print(response.headers)
-    print(response.text)
-    ct = response.headers["x-garage-causality-token"]
-
-    print("-- Delete")
-    response = requests.delete('http://localhost:3812/alex/root?sort_key=%s'%sk,
-                            headers={'x-garage-causality-token': ct},
-                            auth=auth)
-    print(response.headers)
-    print(response.text)
-
-print("-- ReadIndex")
-response = requests.get('http://localhost:3812/alex',
-                        auth=auth)
-print(response.headers)
-print(response.text)
-
-print("-- InsertBatch")
-response = requests.post('http://localhost:3812/alex',
-                        auth=auth,
-                        data='''
-[
-    {"pk": "root", "sk": "a", "ct": null, "v": "aW5pdGlhbCB0ZXN0Cg=="},
-    {"pk": "root", "sk": "b", "ct": null, "v": "aW5pdGlhbCB0ZXN1Cg=="},
-    {"pk": "root", "sk": "c", "ct": null, "v": "aW5pdGlhbCB0ZXN2Cg=="}
-]
-''')
-print(response.headers)
-print(response.text)
-
-print("-- ReadIndex")
-response = requests.get('http://localhost:3812/alex',
-                        auth=auth)
-print(response.headers)
-print(response.text)
-
-for sk in sort_keys:
-    print("-- (%s) Get"%sk)
-    response = requests.get('http://localhost:3812/alex/root?sort_key=%s'%sk,
-                            auth=auth)
-    print(response.headers)
-    print(response.text)
-    ct = response.headers["x-garage-causality-token"]
-
-print("-- ReadBatch")
-response = requests.post('http://localhost:3812/alex?search',
-                        auth=auth,
-                        data='''
-[
-    {"partitionKey": "root"},
-    {"partitionKey": "root", "tombstones": true},
-    {"partitionKey": "root", "tombstones": true, "limit": 2},
-    {"partitionKey": "root", "start": "c", "singleItem": true},
-    {"partitionKey": "root", "start": "b", "end": "d", "tombstones": true}
-]
-''')
-print(response.headers)
-print(response.text)
-
-
-print("-- DeleteBatch")
-response = requests.post('http://localhost:3812/alex?delete',
-                        auth=auth,
-                        data='''
-[
-    {"partitionKey": "root", "start": "b", "end": "c"}
-]
-''')
-print(response.headers)
-print(response.text)
-
-print("-- ReadBatch")
-response = requests.post('http://localhost:3812/alex?search',
-                        auth=auth,
-                        data='''
-[
-    {"partitionKey": "root"}
-]
-''')
-print(response.headers)
-print(response.text)

nix/compile.nix

@@ -173,6 +173,7 @@ let
       "garage/kubernetes-discovery"
       "garage/metrics"
       "garage/telemetry-otlp"
+      "garage/syslog"
     ] else
       [ ]));
 

src/block/layout.rs

@@ -1,3 +1,4 @@
+use std::collections::HashMap;
 use std::path::PathBuf;
 
 use serde::{Deserialize, Serialize};
@@ -13,9 +14,12 @@ const DRIVE_NPART: usize = 1024;
 
 const HASH_DRIVE_BYTES: (usize, usize) = (2, 3);
 
+const MARKER_FILE_NAME: &str = "garage-marker";
+
 #[derive(Serialize, Deserialize, Debug, Clone)]
 pub(crate) struct DataLayout {
 	pub(crate) data_dirs: Vec<DataDir>,
+	markers: HashMap<PathBuf, String>,
 
 	/// Primary storage location (index in data_dirs) for each partition
 	/// = the location where the data is supposed to be, blocks are always
@@ -75,16 +79,17 @@ impl DataLayout {
 
 		Ok(Self {
 			data_dirs,
+			markers: HashMap::new(),
 			part_prim,
 			part_sec,
 		})
 	}
 
-	pub(crate) fn update(&mut self, dirs: &DataDirEnum) -> Result<(), Error> {
+	pub(crate) fn update(self, dirs: &DataDirEnum) -> Result<Self, Error> {
 		// Make list of new data directories, exit if nothing changed
 		let data_dirs = make_data_dirs(dirs)?;
 		if data_dirs == self.data_dirs {
-			return Ok(());
+			return Ok(self);
 		}
 
 		let total_cap = data_dirs.iter().filter_map(|x| x.capacity()).sum::<u64>();
@@ -214,11 +219,43 @@ impl DataLayout {
 		}
 
 		// Apply newly generated config
-		*self = Self {
+		Ok(Self {
 			data_dirs,
+			markers: self.markers,
 			part_prim,
 			part_sec,
-		};
+		})
+	}
+
+	pub(crate) fn check_markers(&mut self) -> Result<(), Error> {
+		let data_dirs = &self.data_dirs;
+		self.markers
+			.retain(|k, _| data_dirs.iter().any(|x| x.path == *k));
+
+		for dir in self.data_dirs.iter() {
+			let mut marker_path = dir.path.clone();
+			marker_path.push(MARKER_FILE_NAME);
+			let existing_marker = std::fs::read_to_string(&marker_path).ok();
+			match (existing_marker, self.markers.get(&dir.path)) {
+				(Some(m1), Some(m2)) => {
+					if m1 != *m2 {
+						return Err(Error::Message(format!("Mismatched content for marker file `{}` in data directory `{}`. If you moved data directories or changed their mountpoints, you should remove the `data_layout` file in Garage's metadata directory and restart Garage.", MARKER_FILE_NAME, dir.path.display())));
+					}
+				}
+				(None, Some(_)) => {
+					return Err(Error::Message(format!("Could not find expected marker file `{}` in data directory `{}`, make sure this data directory is mounted correctly.", MARKER_FILE_NAME, dir.path.display())));
+				}
+				(Some(mkr), None) => {
+					self.markers.insert(dir.path.clone(), mkr);
+				}
+				(None, None) => {
+					let mkr = hex::encode(garage_util::data::gen_uuid().as_slice());
+					std::fs::write(&marker_path, &mkr)?;
+					self.markers.insert(dir.path.clone(), mkr);
+				}
+			}
+		}
+
 		Ok(())
 	}
 
@@ -255,6 +292,7 @@ impl DataLayout {
 	pub(crate) fn without_secondary_locations(&self) -> Self {
 		Self {
 			data_dirs: self.data_dirs.clone(),
+			markers: self.markers.clone(),
 			part_prim: self.part_prim.clone(),
 			part_sec: self.part_sec.iter().map(|_| vec![]).collect::<Vec<_>>(),
 		}
@@ -322,14 +360,12 @@ fn make_data_dirs(dirs: &DataDirEnum) -> Result<Vec<DataDir>, Error> {
 fn dir_not_empty(path: &PathBuf) -> Result<bool, Error> {
 	for entry in std::fs::read_dir(&path)? {
 		let dir = entry?;
-		if dir.file_type()?.is_dir()
-			&& dir
-				.file_name()
-				.into_string()
-				.ok()
-				.and_then(|hex| hex::decode(&hex).ok())
-				.is_some()
-		{
+		let ft = dir.file_type()?;
+		let name = dir.file_name().into_string().ok();
+		if ft.is_file() && name.as_deref() == Some(MARKER_FILE_NAME) {
+			return Ok(true);
+		}
+		if ft.is_dir() && name.and_then(|hex| hex::decode(&hex).ok()).is_some() {
 			return Ok(true);
 		}
 	}

src/block/manager.rs

@@ -127,17 +127,15 @@ impl BlockManager {
 		// Load or compute layout, i.e. assignment of data blocks to the different data directories
 		let data_layout_persister: Persister<DataLayout> =
 			Persister::new(&system.metadata_dir, "data_layout");
-		let data_layout = match data_layout_persister.load() {
-			Ok(mut layout) => {
-				layout
-					.update(&config.data_dir)
-					.ok_or_message("invalid data_dir config")?;
-				layout
-			}
+		let mut data_layout = match data_layout_persister.load() {
+			Ok(layout) => layout
+				.update(&config.data_dir)
+				.ok_or_message("invalid data_dir config")?,
 			Err(_) => {
 				DataLayout::initialize(&config.data_dir).ok_or_message("invalid data_dir config")?
 			}
 		};
+		data_layout.check_markers()?;
 		data_layout_persister
 			.save(&data_layout)
 			.expect("cannot save data_layout");

src/garage/Cargo.toml

@@ -59,6 +59,7 @@ opentelemetry.workspace = true
 opentelemetry-prometheus = { workspace = true, optional = true }
 opentelemetry-otlp = { workspace = true, optional = true }
 prometheus = { workspace = true, optional = true }
+syslog-tracing = { workspace = true, optional = true }
 
 [dev-dependencies]
 aws-config.workspace = true
@@ -97,6 +98,8 @@ kubernetes-discovery = [ "garage_rpc/kubernetes-discovery" ]
 metrics = [ "garage_api/metrics", "opentelemetry-prometheus", "prometheus" ]
 # Exporter for the OpenTelemetry Collector.
 telemetry-otlp = [ "opentelemetry-otlp" ]
+# Logging to syslog
+syslog = [ "syslog-tracing" ]
 
 # NOTE: bundled-libs and system-libs should be treat as mutually exclusive;
 # exactly one of them should be enabled.

src/garage/main.rs

@@ -138,17 +138,8 @@ async fn main() {
 	let opt = Opt::from_clap(&Opt::clap().version(version.as_str()).get_matches());
 
 	// Initialize logging as well as other libraries used in Garage
-	if std::env::var("RUST_LOG").is_err() {
-		let default_log = match &opt.cmd {
-			Command::Server => "netapp=info,garage=info",
-			_ => "netapp=warn,garage=warn",
-		};
-		std::env::set_var("RUST_LOG", default_log)
-	}
-	tracing_subscriber::fmt()
-		.with_writer(std::io::stderr)
-		.with_env_filter(tracing_subscriber::filter::EnvFilter::from_default_env())
-		.init();
+	init_logging(&opt);
+
 	sodiumoxide::init().expect("Unable to init sodiumoxide");
 
 	let res = match opt.cmd {
@@ -171,6 +162,58 @@ async fn main() {
 	}
 }
 
+fn init_logging(opt: &Opt) {
+	if std::env::var("RUST_LOG").is_err() {
+		let default_log = match &opt.cmd {
+			Command::Server => "netapp=info,garage=info",
+			_ => "netapp=warn,garage=warn",
+		};
+		std::env::set_var("RUST_LOG", default_log)
+	}
+
+	let env_filter = tracing_subscriber::filter::EnvFilter::from_default_env();
+
+	if std::env::var("GARAGE_LOG_TO_SYSLOG")
+		.map(|x| x == "1" || x == "true")
+		.unwrap_or(false)
+	{
+		#[cfg(feature = "syslog")]
+		{
+			use std::ffi::CStr;
+			use syslog_tracing::{Facility, Options, Syslog};
+
+			let syslog = Syslog::new(
+				CStr::from_bytes_with_nul(b"garage\0").unwrap(),
+				Options::LOG_PID | Options::LOG_PERROR,
+				Facility::Daemon,
+			)
+			.expect("Unable to init syslog");
+
+			tracing_subscriber::fmt()
+				.with_writer(syslog)
+				.with_env_filter(env_filter)
+				.with_ansi(false) // disable ANSI escape sequences (colours)
+				.with_file(false)
+				.with_level(false)
+				.without_time()
+				.compact()
+				.init();
+
+			return;
+		}
+		#[cfg(not(feature = "syslog"))]
+		{
+			eprintln!("Syslog support is not enabled in this build.");
+			std::process::exit(1);
+		}
+	}
+
+	tracing_subscriber::fmt()
+		.with_writer(std::io::stderr)
+		.with_env_filter(env_filter)
+		.init();
+}
+
 async fn cli_command(opt: Opt) -> Result<(), Error> {
 	let config = if (opt.secrets.rpc_secret.is_none() && opt.secrets.rpc_secret_file.is_none())
 		|| opt.rpc_host.is_none()

src/garage/tests/k2v/poll.rs

@@ -10,6 +10,7 @@ use crate::common;
 use crate::json_body;
 
 #[tokio::test]
+#[ignore = "currently broken"]
 async fn test_poll_item() {
 	let ctx = common::context();
 	let bucket = ctx.create_bucket("test-k2v-poll-item");
@@ -98,6 +99,7 @@ async fn test_poll_item() {
 }
 
 #[tokio::test]
+#[ignore = "currently broken"]
 async fn test_poll_range() {
 	let ctx = common::context();
 	let bucket = ctx.create_bucket("test-k2v-poll-range");