Cassandra from memory heap
I have 4 cassandra nodes in a cluster and one column family that has 10 columns where the row cannot grow very wide (maybe 1000 columns).
I have a "peak" record where I insert up to 500,000 records in a range of 5-10 minutes.
I am using node.js driver: node-cassandra-cql
3 nodes work fine, but one node crashes every time on heavy write.
Currently all nodes are around 1.5 GB in size and the problem node has a data size of 1.9 GB.
All nodes have a maximum heap space of 1 GB (I have 4 GB of RAM on the machines, so the cassandra config file calculated this heap amount by default)
I am using the default cassandra config, except I increased the write / read timeouts.
Question : Does anyone know what could be causing this? Is the heap size really that small? What and how to set up a cassandra cluster for this use case (heavy writing on a small time interval and at other times actually does nothing, or just a small write)
I have not tried to increase the heap size manually, at first I would like to know, maybe there is something else to tweak instead, just increasing it.
source to share
No one has answered this question yet
Check out similar questions: