Submitting forms on the support site are temporary unavailable for schedule maintenance. If you need immediate assistance please contact technical support. We apologize for the inconvenience.
What is common terminology used with the Cassandra agent?
설명
What common terms are used in Cassandra database environments?
해결 방안
Cassandra – An open source, distributed, highly available and partition tolerant database that runs on the JVM.
Cluster - Also called a ring, this is a collection of intercommunicating servers that together hold one or more distributed databases. Since Cassandra is designed for the purposes of throughput, availability, and data distribution, it is never run as a single instance: A cluster will, in practice, always contain multiple nodes.
Node - A single database server.
Coordinator - The node originally contacted with a query. It is responsible for assembling and returning the response to the client, but may or may not house any of the queried data and is in no way different from any other node in the cluster. Any node can and will act as a coordinator for any given query, though most clients are designed to be efficient by, all else being equal, contacting the 'closest' node in the cluster (e.g. by lowest latency).
JMX - A protocol for exposing data and remote method invocation hooks from within a Java application.
Keyspace - The Cassandra terminology for a database/schema.
CQL - The Cassandra Query Language was designed to be very similar in format to SQL.
cqlsh – The command-line CQL shell application, distributed with Cassandra.
nodetool – A command-line Cassandra management utility. Can be used to check up/down status of nodes, change runtime settings, and run a wide range of diagnostic commands. • Table - Corresponds roughly with a table from a relational database. The internal data model is different from a true relational table and this can sometimes be relevant for the application developer and schema design considerations. Formerly called "Column family".
Partition key - The column on which data is partitioned across nodes in a cluster.
Clustering columns - Columns that enable differentiation of rows that have the same partition key value. A table definition may have zero or more clustering columns.
Partition - One or more rows that all have identical values for the partition key. A partition is not split across nodes (though it is likely replicated in its entirety to multiple nodes).
Row - Like its namesake in a relational database.
Apache Cassandra vs. DSE - Cassandra started as an open source project at Facebook, based off of the data model of Google's Big Table. It was accepted as a top level project at the Apache Software Foundation in 2010. The private company DataStax drives much of the development of Cassandra and has forked the project to create a propriety, though largely compatible, closed source solution called DataStax Enterprise, or DSE. DSE has been based on older versions of Apache Cassandra, but is packaged with novel features and utilities and is offered with support plans. DSE is offered with a product called OpsCenter that provides limited monitoring capabilities. It requires that an agent be installed on each node, in contrast to the Foglight agent which will usually be set up to simultaneously monitor several nodes remotely.
추가 정보
Source: Foglight for Cassandra Installation and Troubleshooting guide.