반응형
Gossip protocol
- Gossip is a peer-to-peer communication protocol in which nodes periodically exchanges state information about themselves and about other nodes they know about.
- the gossip process runs every second and exchanges state messages with up to three other nodes in the cluster.
Seed nodes
- list of seed nodes
- has no purpose other than bootstraping thre gossip process for new nodes joining the cluster.
- seed nodes are not a SPOF, nor do they have any other special purpose in cluster operations beyond thre bootstraping of nodes.
- in multiple data-center clusters, include at least one node from each datacenter in three seed list.
- desigating more than a single seed node per datacenter is recommended for fault tolerance.
- otherwise gossip has to communicate with another datacenter when bootstraping a node.
- making every node a seed node is not recommended because of increased maintenance and reduced gossip performance.
- recommended to use approximately three nodes per datacenter.
반응형
'DBMS > Cassandra' 카테고리의 다른 글
Cassandra Data replication (0) | 2023.03.03 |
---|---|
Cassandra Partitioners (0) | 2023.03.03 |
Cassandra 세컨더리 인덱스의 위험 이유 및 대응 방안 (0) | 2022.12.18 |
카산드라 Batch operation (0) | 2022.12.18 |
[Cassandra] 인증/권한 설정하기 (0) | 2022.09.30 |