Hello Dojo Community,
Recently, one of our internal DMSs used for testing was hit by a Storage RAID failure at HQ. We have rebuilt Cassandra and Elasticsearch nodes, and everything is functioning correctly again.
However, during the database rebuilding and DMS connection process, I noticed something I didn't fully understand. According to the documentation, when configuring a general database of type 'CassandraCluster' (i.e., one Cassandra cluster for the entire DMS), it states that this configuration also requires the setup of an Indexing database.
Indeed, when I brought up the Cassandra nodes and tried to connect them to the DMS, the connection couldn't be established. It wasn't until I brought up the Elasticsearch nodes and connected both Cassandra cluster and Elasticsearch to the DMS that the connection was successful.
I am now curious whether it is possible to have just a general database of type 'CassandraCluster' without an Indexing DB. If that isn't the case, what is the reasoning behind it, and why can't the DMS simply use 'CassandraCluster' as the sole database if an Indexing DB is not needed?
Thank you.
Hi,
When using CassandraCluster and indexing database is required.It is not possible to configure the database without one as both DataBases holds unique sets of data that are not duplicated in the other type of DataBase.
The reason for this is that both databases have their strength and weaknesses. The indexing DB stores data where requests are more defined at runtime, such as looking up alarms based on certain attributes depending on the users need.
The Cassandra holds data where the queries are more rigid, such as element data where the queries are known to request data on an element or parameter level.
Kind regards,
Yeah, really interesting question and answer.
Thank you, Brent, for the concise and clear explanation.
Wkr.