Hi,
We have deployed cassandra using k8ssandra operator with 1 DC and 3 racks(each rack in a different zone). We have installed temporal application on this cassandra DB. The keyspace temporal has replica as 3(1 per node). We also have reaper also installed as part of k8ssandra operator and an incremental repair run on a daily basis.
Current size of the nodes is around 430 GB on each node and i have executed garbagecollect manually and have noticed the node size reduced to 40 gb from 430 gb.
How do we verify the garbagecollect is being carried out frequently or is there an automated way of running garbagecollection using reaper?
I have got the same reduction in size by running nodetool compact as well. Size is reduced from 430gb to 40gb. Definition of table is set to level compaction strategy, so it should ideally autocompact.
Is it advised to initiate nodetool compact manually?
note: The gc_grace_seconds value is set to default 864000