I set up Kubernetes for a limit of 64GB RAM for a Scylla pod. What if Scylla hits 64GB? Will my system crash? Or will it auto-free memory?
Also, if I increase the memory limit to 96GB, will this improve performance?
I set up Kubernetes for a limit of 64GB RAM for a Scylla pod. What if Scylla hits 64GB? Will my system crash? Or will it auto-free memory?
Also, if I increase the memory limit to 96GB, will this improve performance?
ScyllaDB grabs all available memory and manages it internally.
It is completely normal to have ScyllaDB at the memory limit. This is by design.
Regarding increasing the memory limit, of course, the more memory, the better. The excess memory – that is not strictly needed for normal operations – is used to cache data from the disk, greatly increasing access speed for cached rows. More memory also allows for supporting more storage (ScyllaDB aims for a 1:100 memory:disk ratio).