This document describes the differences between Database clustered and Database connected on a Cisco Meeting Server (CMS).
In order to create a Callbridge cluster, all the Callbridge part of the cluster must share the same Database, there are two methods to share the same Database, cluster the Database service or Connect to a Database already created. In this document you can see the difference between both methods, limitations and advantages.
Contributed by Amadeus Ubaldo and Octavio Miralrio, Cisco TAC Engineers.
What are the differences between a Database clustered and Database connected?
- A Database cluster is an online back of the Database which is maintained while the system runs
- database cluster initialize and database cluster join commands are used for this purpose
- On the Database cluster creation, it is required to select a primary Database and then, replicate it to the rest of the peers on the Database cluster
- An odd number of peers on the Database cluster are required, since in case of a failure, the process to select a new primary Database and continue with the services uninterrupted, and a maximum of five Database servers
- Database cluster does not have the capabilties to perform any kind of load balance or cache the information. Additionally, it does not perform any shard of the data in order to make the local access more efficient within any kind of geographically distributed arrangement.
- All queries are directed at the Database which serves as the primary node, what ever it is
- The replicas are not available as read-only instances
- Escentialy, when a cluster of Callbridge is needed, all the Callbridges need to be connected to the same Database
- When Database is not clustered, one Database server is used to handle all the Callbridges, that means all Callbridges in the cluster need to connecte to the same Database
- For this purpose, the command database cluster initialize needs to be run on the primary and only Database server and database cluster connect is used on the Callbridge servers that are part of the Callbridge cluster
- This allows to avoid configure a Database service on every Callbridge part of the cluster
- And also avoid the limitation on the odd number of Database servers
- The limitation on Database connected is that if the only node where the Database is enabled goes down, the entire deployment goes down
- The connection applies also when a Database cluster already exists with already five servers configured, but more than five Callbridges are needed, the new Callbridge peers can be onlu connected and act as part of the Callbridge cluster but no the Database cluster.
Callbridge Database connect to Database cluster:
- Callbridges which are already allocated with a database (primary or replica) are automatically connected to the same database cluster that the co-allocated database is already part of
- The term connect, means that the Callbridge knows how to access all the Databases in a cluster
- Unlike the attach command used in the Database cluster creation, the connect command does not delete any database on the server that host Callbridge. Therefore if you use the command on a Core server with a local Database then the contents of the Database are not destroyed, but the local Database is invisible until this server’s Call Bridge is disconnected from the cluster