Reputation: 187
So currently I have an application which uses cassandra. I have 3 cassandra nodes of which 1 is seed node. The app currently takes in around 100 write requests every 20 seconds and around 200 read requests every second. The app seems to be crashing a lot with the error - expected to return 1 entry, but received 0.
Just wondering what optimization steps I should take into an account, so the cassandra does not crash that much anymore?
I'm following the correct table structure, but I'm not using materialized views, instead I'm using ALLOW FILTERING - could that be the problem?
Are there any other suggestions?
Upvotes: 1
Views: 124
Reputation: 16430
ALLOW FILTERING
is there strictly for testing and dev, occasional ops purposes. While some tools like spark have been optimized with it, it should never be used by your application. There is no way to make it work efficiently or optimizations for it, please do not use it. If a query requires it, you have a data modeling problem. Your tables should reflect your queries not your data.
Upvotes: 4