Many improvements on ring/replication and its configuration:

- Explicit "replication_mode" configuration parameters that takes
  either "none", "2" or "3" as values, instead of letting user configure
  replication factor themselves. These are presets whose corresponding
  replication/quorum values can be found in replication/mode.rs

- Explicit support for single-node and two-node deployments
  (number of nodes must be at least "replication_mode", with "none"
  we can have only one node)

- Ring is now stored much more compactly with 256*8 + n*32 bytes,
  instead of 256*32 bytes

- Support for gateway-only nodes that do not store data
  (these nodes still need a metadata_directory to store the list
  of bucket and keys since those are stored on all nodes; it also
  technically needs a data_directory to start but it will stay
  empty unless we have bugs)
This commit is contained in:
Alex Auvolat 2021-05-28 12:36:22 +02:00
parent c8aa1eb481
commit b490ebc7f6
No known key found for this signature in database
GPG key ID: EDABF9711E244EB1
10 changed files with 252 additions and 109 deletions

View file

@ -31,9 +31,7 @@ bootstrap_peers = [
"127.0.0.1:3903"
]
max_concurrent_rpc_requests = 12
data_replication_factor = 3
meta_replication_factor = 3
meta_epidemic_fanout = 3
replication_mode = "3"
[s3_api]
api_bind_addr = "0.0.0.0:$((3910+$count))" # the S3 API port, HTTP without TLS. Add a reverse proxy for the TLS part.

View file

@ -430,8 +430,8 @@ impl AdminRpcHandler {
// Gather ring statistics
let ring = self.garage.system.ring.borrow().clone();
let mut ring_nodes = HashMap::new();
for r in ring.ring.iter() {
for n in r.nodes.iter() {
for (_i, loc) in ring.partitions().iter() {
for n in ring.get_nodes(loc, ring.replication_factor).iter() {
if !ring_nodes.contains_key(n) {
ring_nodes.insert(*n, 0usize);
}

View file

@ -80,6 +80,10 @@ pub struct ConfigureNodeOpt {
#[structopt(short = "c", long = "capacity")]
capacity: Option<u32>,
/// Gateway-only node
#[structopt(short = "g", long = "gateway")]
gateway: bool,
/// Optional node tag
#[structopt(short = "t", long = "tag")]
tag: Option<String>,
@ -339,7 +343,12 @@ pub async fn cmd_status(
if let Some(cfg) = config.members.get(&adv.id) {
println!(
"{:?}\t{}\t{}\t[{}]\t{}\t{}",
adv.id, adv.state_info.hostname, adv.addr, cfg.tag, cfg.datacenter, cfg.capacity
adv.id,
adv.state_info.hostname,
adv.addr,
cfg.tag,
cfg.datacenter,
cfg.capacity_string()
);
} else {
println!(
@ -366,7 +375,7 @@ pub async fn cmd_status(
adv.addr,
cfg.tag,
cfg.datacenter,
cfg.capacity,
cfg.capacity_string(),
(now_msec() - adv.last_seen) / 1000,
);
}
@ -375,7 +384,10 @@ pub async fn cmd_status(
if !status.iter().any(|x| x.id == *id) {
println!(
"{:?}\t{}\t{}\t{}\tnever seen",
id, cfg.tag, cfg.datacenter, cfg.capacity
id,
cfg.tag,
cfg.datacenter,
cfg.capacity_string(),
);
}
}
@ -438,23 +450,44 @@ pub async fn cmd_configure(
}
}
if args.capacity.is_some() && args.gateway {
return Err(Error::Message(
"-c and -g are mutually exclusive, please configure node either with c>0 to act as a storage node or with -g to act as a gateway node".into()));
}
if args.capacity == Some(0) {
return Err(Error::Message("Invalid capacity value: 0".into()));
}
let new_entry = match config.members.get(&added_node) {
None => NetworkConfigEntry {
datacenter: args
.datacenter
.expect("Please specifiy a datacenter with the -d flag"),
capacity: args
.capacity
.expect("Please specifiy a capacity with the -c flag"),
tag: args.tag.unwrap_or_default(),
},
Some(old) => NetworkConfigEntry {
datacenter: args
.datacenter
.unwrap_or_else(|| old.datacenter.to_string()),
capacity: args.capacity.unwrap_or(old.capacity),
tag: args.tag.unwrap_or_else(|| old.tag.to_string()),
},
None => {
let capacity = match args.capacity {
Some(c) => Some(c),
None if args.gateway => None,
_ => return Err(Error::Message(
"Please specify a capacity with the -c flag, or set node explicitly as gateway with -g".into())),
};
NetworkConfigEntry {
datacenter: args
.datacenter
.expect("Please specifiy a datacenter with the -d flag"),
capacity,
tag: args.tag.unwrap_or_default(),
}
}
Some(old) => {
let capacity = match args.capacity {
Some(c) => Some(c),
None if args.gateway => None,
_ => old.capacity,
};
NetworkConfigEntry {
datacenter: args
.datacenter
.unwrap_or_else(|| old.datacenter.to_string()),
capacity,
tag: args.tag.unwrap_or_else(|| old.tag.to_string()),
}
}
};
config.members.insert(added_node, new_entry);

View file

@ -7,6 +7,7 @@ use garage_rpc::membership::System;
use garage_rpc::rpc_client::RpcHttpClient;
use garage_rpc::rpc_server::RpcServer;
use garage_table::replication::ReplicationMode;
use garage_table::replication::TableFullReplication;
use garage_table::replication::TableShardedReplication;
use garage_table::*;
@ -50,6 +51,9 @@ impl Garage {
background: Arc<BackgroundRunner>,
rpc_server: &mut RpcServer,
) -> Arc<Self> {
let replication_mode = ReplicationMode::parse(&config.replication_mode)
.expect("Invalid replication_mode in config file.");
info!("Initialize membership management system...");
let rpc_http_client = Arc::new(
RpcHttpClient::new(config.max_concurrent_rpc_requests, &config.rpc_tls)
@ -60,32 +64,33 @@ impl Garage {
rpc_http_client,
background.clone(),
rpc_server,
replication_mode.replication_factor(),
);
let data_rep_param = TableShardedReplication {
system: system.clone(),
replication_factor: config.data_replication_factor,
write_quorum: (config.data_replication_factor + 1) / 2,
replication_factor: replication_mode.replication_factor(),
write_quorum: replication_mode.write_quorum(),
read_quorum: 1,
};
let meta_rep_param = TableShardedReplication {
system: system.clone(),
replication_factor: config.meta_replication_factor,
write_quorum: (config.meta_replication_factor + 1) / 2,
read_quorum: (config.meta_replication_factor + 1) / 2,
replication_factor: replication_mode.replication_factor(),
write_quorum: replication_mode.write_quorum(),
read_quorum: replication_mode.read_quorum(),
};
let control_rep_param = TableFullReplication {
system: system.clone(),
max_faults: config.control_write_max_faults,
max_faults: replication_mode.control_write_max_faults(),
};
info!("Initialize block manager...");
let block_manager = BlockManager::new(
&db,
config.data_dir.clone(),
data_rep_param.clone(),
data_rep_param,
system.clone(),
rpc_server,
);
@ -95,7 +100,7 @@ impl Garage {
BlockRefTable {
block_manager: block_manager.clone(),
},
data_rep_param,
meta_rep_param.clone(),
system.clone(),
&db,
"block_ref".to_string(),

View file

@ -95,6 +95,7 @@ pub struct System {
rpc_http_client: Arc<RpcHttpClient>,
rpc_client: Arc<RpcClient<Message>>,
replication_factor: usize,
pub(crate) status: watch::Receiver<Arc<Status>>,
/// The ring
pub ring: watch::Receiver<Arc<Ring>>,
@ -228,6 +229,7 @@ impl System {
rpc_http_client: Arc<RpcHttpClient>,
background: Arc<BackgroundRunner>,
rpc_server: &mut RpcServer,
replication_factor: usize,
) -> Arc<Self> {
let id = gen_node_id(&metadata_dir).expect("Unable to read or generate node ID");
info!("Node ID: {}", hex::encode(&id));
@ -259,7 +261,7 @@ impl System {
.unwrap_or_else(|_| "<invalid utf-8>".to_string()),
};
let ring = Ring::new(net_config);
let ring = Ring::new(net_config, replication_factor);
let (update_ring, ring) = watch::channel(Arc::new(ring));
let rpc_path = MEMBERSHIP_RPC_PATH.to_string();
@ -277,6 +279,7 @@ impl System {
state_info,
rpc_http_client,
rpc_client,
replication_factor,
status,
ring,
update_lock: Mutex::new(Updaters {
@ -543,7 +546,7 @@ impl System {
let ring: Arc<Ring> = self.ring.borrow().clone();
if adv.version > ring.config.version {
let ring = Ring::new(adv.clone());
let ring = Ring::new(adv.clone(), self.replication_factor);
update_lock.update_ring.send(Arc::new(ring))?;
drop(update_lock);

View file

@ -7,10 +7,9 @@ use serde::{Deserialize, Serialize};
use garage_util::data::*;
// A partition number is encoded on 16 bits,
// i.e. we have up to 2**16 partitions.
// (in practice we have exactly 2**PARTITION_BITS partitions)
/// A partition id, stored on 16 bits
/// A partition id, which is stored on 16 bits
/// i.e. we have up to 2**16 partitions.
/// (in practice we have exactly 2**PARTITION_BITS partitions)
pub type Partition = u16;
// TODO: make this constant parametrizable in the config file
@ -23,11 +22,6 @@ pub const PARTITION_BITS: usize = 8;
const PARTITION_MASK_U16: u16 = ((1 << PARTITION_BITS) - 1) << (16 - PARTITION_BITS);
// TODO: make this constant paraetrizable in the config file
// (most deployments use a replication factor of 3, so...)
/// The maximum number of time an object might get replicated
pub const MAX_REPLICATION: usize = 3;
/// The user-defined configuration of the cluster's nodes
#[derive(Clone, Debug, Serialize, Deserialize)]
pub struct NetworkConfig {
@ -53,40 +47,72 @@ pub struct NetworkConfigEntry {
/// geodistribution
pub datacenter: String,
/// The (relative) capacity of the node
pub capacity: u32,
/// If this is set to None, the node does not participate in storing data for the system
/// and is only active as an API gateway to other nodes
pub capacity: Option<u32>,
/// A tag to recognize the entry, not used for other things than display
pub tag: String,
}
impl NetworkConfigEntry {
pub fn capacity_string(&self) -> String {
match self.capacity {
Some(c) => format!("{}", c),
None => "gateway".to_string(),
}
}
}
/// A ring distributing fairly objects to nodes
#[derive(Clone)]
pub struct Ring {
/// The replication factor for this ring
pub replication_factor: usize,
/// The network configuration used to generate this ring
pub config: NetworkConfig,
/// The list of entries in the ring
pub ring: Vec<RingEntry>,
// Internal order of nodes used to make a more compact representation of the ring
nodes: Vec<Uuid>,
// The list of entries in the ring
ring: Vec<RingEntry>,
}
// Type to store compactly the id of a node in the system
// Change this to u16 the day we want to have more than 256 nodes in a cluster
type CompactNodeType = u8;
// The maximum number of times an object might get replicated
// This must be at least 3 because Garage supports 3-way replication
// Here we use 6 so that the size of a ring entry is 8 bytes
// (2 bytes partition id, 6 bytes node numbers as u8s)
const MAX_REPLICATION: usize = 6;
/// An entry in the ring
#[derive(Clone, Debug)]
pub struct RingEntry {
/// The prefix of the Hash of object which should use this entry
pub location: Hash,
/// The nodes in which a matching object should get stored
pub nodes: [Uuid; MAX_REPLICATION],
struct RingEntry {
// The two first bytes of the first hash that goes in this partition
// (the next bytes are zeroes)
hash_prefix: u16,
// The nodes that store this partition, stored as a list of positions in the `nodes`
// field of the Ring structure
// Only items 0 up to ring.replication_factor - 1 are used, others are zeros
nodes_buf: [CompactNodeType; MAX_REPLICATION],
}
impl Ring {
// TODO this function MUST be refactored, it's 100 lines long, with a 50 lines loop, going up to 6
// levels of imbrication. It is basically impossible to test, maintain, or understand for an
// outsider.
pub(crate) fn new(config: NetworkConfig) -> Self {
pub(crate) fn new(config: NetworkConfig, replication_factor: usize) -> Self {
// Create a vector of partition indices (0 to 2**PARTITION_BITS-1)
let partitions_idx = (0usize..(1usize << PARTITION_BITS)).collect::<Vec<_>>();
let datacenters = config
.members
.iter()
.filter(|(_id, info)| info.capacity.is_some())
.map(|(_id, info)| info.datacenter.as_str())
.collect::<HashSet<&str>>();
let n_datacenters = datacenters.len();
@ -101,6 +127,7 @@ impl Ring {
let mut queues = config
.members
.iter()
.filter(|(_id, info)| info.capacity.is_some())
.map(|(node_id, node_info)| {
let mut parts = partitions_idx
.iter()
@ -119,11 +146,13 @@ impl Ring {
let max_capacity = config
.members
.iter()
.map(|(_, node_info)| node_info.capacity)
.filter_map(|(_, node_info)| node_info.capacity)
.fold(0, std::cmp::max);
assert!(replication_factor <= MAX_REPLICATION);
// Fill up ring
for rep in 0..MAX_REPLICATION {
for rep in 0..replication_factor {
queues.sort_by_key(|(ni, _np, _q, _p)| {
let queue_data = [&u16::to_be_bytes(rep as u16)[..], ni.as_slice()].concat();
fasthash(&queue_data[..])
@ -138,7 +167,7 @@ impl Ring {
let remaining0 = remaining;
for i_round in 0..max_capacity {
for (node_id, node_info, q, pos) in queues.iter_mut() {
if i_round >= node_info.capacity {
if i_round >= node_info.capacity.unwrap() {
continue;
}
for (pos2, &qv) in q.iter().enumerate().skip(*pos) {
@ -166,34 +195,58 @@ impl Ring {
// No progress made, exit
warn!("Could not build ring, not enough nodes configured.");
return Self {
replication_factor,
config,
nodes: vec![],
ring: vec![],
};
}
}
}
// Make a canonical order for nodes
let nodes = config
.members
.iter()
.filter(|(_id, info)| info.capacity.is_some())
.map(|(id, _)| *id)
.collect::<Vec<_>>();
let nodes_rev = nodes
.iter()
.enumerate()
.map(|(i, id)| (*id, i as CompactNodeType))
.collect::<HashMap<Uuid, CompactNodeType>>();
let ring = partitions
.iter()
.enumerate()
.map(|(i, nodes)| {
let top = (i as u16) << (16 - PARTITION_BITS);
let mut hash = [0u8; 32];
hash[0..2].copy_from_slice(&u16::to_be_bytes(top)[..]);
let nodes = nodes.iter().map(|(id, _info)| **id).collect::<Vec<Uuid>>();
let nodes = nodes
.iter()
.map(|(id, _info)| *nodes_rev.get(id).unwrap())
.collect::<Vec<CompactNodeType>>();
assert!(nodes.len() == replication_factor);
let mut nodes_buf = [0u8; MAX_REPLICATION];
nodes_buf[..replication_factor].copy_from_slice(&nodes[..]);
RingEntry {
location: hash.into(),
nodes: nodes.try_into().unwrap(),
hash_prefix: top,
nodes_buf,
}
})
.collect::<Vec<_>>();
Self { config, ring }
Self {
replication_factor,
config,
nodes,
ring,
}
}
/// Get the partition in which data would fall on
pub fn partition_of(&self, from: &Hash) -> Partition {
let top = u16::from_be_bytes(from.as_slice()[0..2].try_into().unwrap());
pub fn partition_of(&self, position: &Hash) -> Partition {
let top = u16::from_be_bytes(position.as_slice()[0..2].try_into().unwrap());
top >> (16 - PARTITION_BITS)
}
@ -202,7 +255,9 @@ impl Ring {
let mut ret = vec![];
for (i, entry) in self.ring.iter().enumerate() {
ret.push((i as u16, entry.location));
let mut location = [0u8; 32];
location[..2].copy_from_slice(&u16::to_be_bytes(entry.hash_prefix)[..]);
ret.push((i as u16, location.into()));
}
if !ret.is_empty() {
assert_eq!(ret[0].1, [0u8; 32].into());
@ -211,28 +266,38 @@ impl Ring {
ret
}
// TODO rename this function as it no longer walk the ring
/// Walk the ring to find the n servers in which data should be replicated
pub fn walk_ring(&self, from: &Hash, n: usize) -> Vec<Uuid> {
pub fn get_nodes(&self, position: &Hash, n: usize) -> Vec<Uuid> {
if self.ring.len() != 1 << PARTITION_BITS {
warn!("Ring not yet ready, read/writes will be lost!");
return vec![];
}
let top = u16::from_be_bytes(from.as_slice()[0..2].try_into().unwrap());
let partition_idx = (top >> (16 - PARTITION_BITS)) as usize;
// TODO why computing two time in the same way and asserting?
assert_eq!(partition_idx, self.partition_of(from) as usize);
let partition_idx = self.partition_of(position) as usize;
let partition = &self.ring[partition_idx];
let partition_top =
u16::from_be_bytes(partition.location.as_slice()[0..2].try_into().unwrap());
// TODO is this an assertion on the validity of PARTITION_MASK_U16? If so, it should
// probably be a test more than a runtime assertion
assert_eq!(partition_top & PARTITION_MASK_U16, top & PARTITION_MASK_U16);
let top = u16::from_be_bytes(position.as_slice()[0..2].try_into().unwrap());
// Check that we haven't messed up our partition table, i.e. that this partition
// table entrey indeed corresponds to the item we are storing
assert_eq!(
partition.hash_prefix & PARTITION_MASK_U16,
top & PARTITION_MASK_U16
);
assert!(n <= partition.nodes.len());
partition.nodes[..n].to_vec()
assert!(n <= self.replication_factor);
partition.nodes_buf[..n]
.iter()
.map(|i| self.nodes[*i as usize])
.collect::<Vec<_>>()
}
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn test_ring_entry_size() {
assert_eq!(std::mem::size_of::<RingEntry>(), 8);
}
}

View file

@ -1,8 +1,10 @@
mod parameters;
mod fullcopy;
mod mode;
mod sharded;
pub use fullcopy::TableFullReplication;
pub use mode::ReplicationMode;
pub use parameters::*;
pub use sharded::TableShardedReplication;

View file

@ -0,0 +1,47 @@
pub enum ReplicationMode {
None,
TwoWay,
ThreeWay,
}
impl ReplicationMode {
pub fn parse(v: &str) -> Option<Self> {
match v {
"none" | "1" => Some(Self::None),
"2" => Some(Self::TwoWay),
"3" => Some(Self::ThreeWay),
_ => None,
}
}
pub fn control_write_max_faults(&self) -> usize {
match self {
Self::None => 0,
_ => 1,
}
}
pub fn replication_factor(&self) -> usize {
match self {
Self::None => 1,
Self::TwoWay => 2,
Self::ThreeWay => 3,
}
}
pub fn read_quorum(&self) -> usize {
match self {
Self::None => 1,
Self::TwoWay => 1,
Self::ThreeWay => 2,
}
}
pub fn write_quorum(&self) -> usize {
match self {
Self::None => 1,
Self::TwoWay => 2,
Self::ThreeWay => 2,
}
}
}

View file

@ -26,8 +26,8 @@ pub struct TableShardedReplication {
impl TableReplication for TableShardedReplication {
fn read_nodes(&self, hash: &Hash) -> Vec<Uuid> {
let ring = self.system.ring.borrow().clone();
ring.walk_ring(&hash, self.replication_factor)
let ring = self.system.ring.borrow();
ring.get_nodes(&hash, self.replication_factor)
}
fn read_quorum(&self) -> usize {
self.read_quorum
@ -35,7 +35,7 @@ impl TableReplication for TableShardedReplication {
fn write_nodes(&self, hash: &Hash) -> Vec<Uuid> {
let ring = self.system.ring.borrow();
ring.walk_ring(&hash, self.replication_factor)
ring.get_nodes(&hash, self.replication_factor)
}
fn write_quorum(&self) -> usize {
self.write_quorum

View file

@ -15,6 +15,17 @@ pub struct Config {
/// Path where to store data. Can be slower, but need higher volume
pub data_dir: PathBuf,
/// Size of data blocks to save to disk
#[serde(default = "default_block_size")]
pub block_size: usize,
/// Replication mode. Supported values:
/// - none, 1 -> no replication
/// - 2 -> 2-way replication
/// - 3 -> 3-way replication
// (we can add more aliases for this later)
pub replication_mode: String,
/// Address to bind for RPC
pub rpc_bind_addr: SocketAddr,
@ -26,6 +37,13 @@ pub struct Config {
/// Consul service name to use
pub consul_service_name: Option<String>,
/// Configuration for RPC TLS
pub rpc_tls: Option<TlsConfig>,
/// Max number of concurrent RPC request
#[serde(default = "default_max_concurrent_rpc_requests")]
pub max_concurrent_rpc_requests: usize,
/// Sled cache size, in bytes
#[serde(default = "default_sled_cache_capacity")]
pub sled_cache_capacity: u64,
@ -34,28 +52,6 @@ pub struct Config {
#[serde(default = "default_sled_flush_every_ms")]
pub sled_flush_every_ms: u64,
/// Max number of concurrent RPC request
#[serde(default = "default_max_concurrent_rpc_requests")]
pub max_concurrent_rpc_requests: usize,
/// Size of data blocks to save to disk
#[serde(default = "default_block_size")]
pub block_size: usize,
#[serde(default = "default_control_write_max_faults")]
pub control_write_max_faults: usize,
/// How many nodes should hold a copy of meta data
#[serde(default = "default_replication_factor")]
pub meta_replication_factor: usize,
/// How many nodes should hold a copy of data
#[serde(default = "default_replication_factor")]
pub data_replication_factor: usize,
/// Configuration for RPC TLS
pub rpc_tls: Option<TlsConfig>,
/// Configuration for S3 api
pub s3_api: ApiConfig,
@ -106,12 +102,6 @@ fn default_max_concurrent_rpc_requests() -> usize {
fn default_block_size() -> usize {
1048576
}
fn default_replication_factor() -> usize {
3
}
fn default_control_write_max_faults() -> usize {
1
}
/// Read and parse configuration
pub fn read_config(config_file: PathBuf) -> Result<Config, Error> {