aboutsummaryrefslogtreecommitdiff
path: root/src/table/replication/sharded.rs
blob: 793d87fd15dcfb2d2bb7ec3cefb9ece852eb9a78 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
use std::sync::Arc;

use garage_rpc::layout::*;
use garage_rpc::system::System;
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> {
		self.system
			.cluster_layout()
			.current()
			.nodes_of(hash, self.replication_factor)
	}
	fn read_quorum(&self) -> usize {
		self.read_quorum
	}

	fn write_nodes(&self, hash: &Hash) -> Vec<Uuid> {
		self.system
			.cluster_layout()
			.current()
			.nodes_of(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.cluster_layout().current().partition_of(hash)
	}
	fn partitions(&self) -> Vec<(Partition, Hash)> {
		self.system.cluster_layout().current().partitions()
	}
}