Alex
b490ebc7f6
- 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)
54 lines
1.5 KiB
Rust
54 lines
1.5 KiB
Rust
use std::sync::Arc;
|
|
|
|
use garage_rpc::membership::System;
|
|
use garage_rpc::ring::*;
|
|
use garage_util::data::*;
|
|
|
|
use crate::replication::*;
|
|
|
|
/// Sharded replication schema:
|
|
/// - based on the ring of nodes, a certain set of neighbors
|
|
/// store entries, given as a function of the position of the
|
|
/// entry's hash in the ring
|
|
/// - reads are done on all of the nodes that replicate the data
|
|
/// - writes as well
|
|
#[derive(Clone)]
|
|
pub struct TableShardedReplication {
|
|
/// The membership manager of this node
|
|
pub system: Arc<System>,
|
|
/// How many time each data should be replicated
|
|
pub replication_factor: usize,
|
|
/// How many nodes to contact for a read, should be at most `replication_factor`
|
|
pub read_quorum: usize,
|
|
/// How many nodes to contact for a write, should be at most `replication_factor`
|
|
pub write_quorum: usize,
|
|
}
|
|
|
|
impl TableReplication for TableShardedReplication {
|
|
fn read_nodes(&self, hash: &Hash) -> Vec<Uuid> {
|
|
let ring = self.system.ring.borrow();
|
|
ring.get_nodes(&hash, self.replication_factor)
|
|
}
|
|
fn read_quorum(&self) -> usize {
|
|
self.read_quorum
|
|
}
|
|
|
|
fn write_nodes(&self, hash: &Hash) -> Vec<Uuid> {
|
|
let ring = self.system.ring.borrow();
|
|
ring.get_nodes(&hash, self.replication_factor)
|
|
}
|
|
fn write_quorum(&self) -> usize {
|
|
self.write_quorum
|
|
}
|
|
fn max_write_errors(&self) -> usize {
|
|
self.replication_factor - self.write_quorum
|
|
}
|
|
|
|
fn partition_of(&self, hash: &Hash) -> Partition {
|
|
self.system.ring.borrow().partition_of(hash)
|
|
}
|
|
fn partitions(&self) -> Vec<(Partition, Hash)> {
|
|
self.system.ring.borrow().partitions()
|
|
}
|
|
}
|