We are currently using Redis as an operational cache in memory. It works well. The problem is that after restarting Redis, we need to refill it, retrieving data from our persistent storage. This overloads our permanent storage outside its capacity and, therefore, recovery takes a lot of time.
We looked at the options for stopping Redis. The best option (without sacrificing performance) is to use AOF with "appendfsync everysec". But with this option, we can lose the last recent data. This is unacceptable. Using AOF with "appednfsync always" has a significant performance hit.
So, we are evaluating the single node Aerospike. Does this guarantee no data loss in the event of a power failure? In response to the write operation, once Aerospike sends success to the client, the data should never be lost, even if I pull out the power cable of the server machine. As I mentioned above, I believe that Redis can provide this guarantee with the "appednfsync always" option. But we do not consider it, since it has a significant decrease in performance.
If Aerospike can do this, I would like to understand in detail how persistence works in Aerospike. Share some resources explaining the same thing.
We are not looking for a distributed system, because we need strong consistency. Data should not be lost during node failures or split brain scripts.
, , ?