So I have the following setup, I have a 8 servers on a local network, each of them running cassandra. 2 of the servers are seed nodes and rest are just regular nodes. If I run nodetool status, everything shows up correctly and all nodes are UN.
I have a express.js application that is running express-cassandra. It has plenty of tables and materialized views (70 in total), I have to restart the express server multiple times, for the cassandra keyspace to be populated with the tables, otherwise it will just stop populating at some point. After the tables are created, I can access all standard tables just fine, but when I try to access any of the materialized view tables, I keep getting error about unconfigured table - InvalidRequest: Error from server: code=2200 [Invalid query] message="unconfigured table ...
Also, if I run my application, all queries to the materialized views fails.
It seems like the database creation process keeps corrupting (due to the reason I have to restart express server to make sure that all tables are created). If I just start the express.js server and leave it running, it will stop creating tables at around 20 tables (I left it running for around 2 hours without any table created). If I restart it, it will again create few tables, and then again stop. And so on, until I have created all tables.
Also, when I try to do repair on that keyspace (which is created by express-cassandra), I keep getting failure to connect to one of the nodes, although nodetool status shows it's up.
Posting an answer here, just so if anyone else receives similar issue is aware what is causing that.
I had an issue with materialized view name - it was too long. Unfortunately express-cassandra does not provide any information about materialized view length in meaningful way.
Did not try the exact length but it was greater than 70 characters.