3

In a situation where a node keeps going up and down (for example, becomes temporarily unresponsive due to GC), is there a way to limit the the hints stored for the problematic node to prevent build up of hints files?

2 Answers 2

2

It is possible to limit the size of hints for a node that is unresponsive or unreachable by specifying a maximum size per node in cassandra.yaml. For example:

max_hints_size_per_host: 1024MiB

Note that this needs to be set on all nodes and will require a restart to take effect. Cheers!

2

Another good cassandra.yaml setting for limiting hint build-up is the max_hint_window_in_ms. By default, it is set to capture hints for 3 hours of node downtime. But it can be adjusted to suit your needs.

max_hint_window_in_ms: 10800000 # 3 hours

Per Erick's note above, this will also need to set per node and require a restart.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.