AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
How to Initialize a TPS From Standby
It's very challenging for really big databases, like for example, five 10 15 terabytes under some significant load. And why is it so? Because initialization at logical level involves basically a kind of dump restore process which takes long and also affects the health of your source cluster. So how to do it better? This is a recipe. We can create us we can first of all have secondary cluster. I'm going to use patroni terminology here. They are all standby. But one of them is leader because we have cascaded replication.So standby leader is replicating using stimuli replication from main primary. And then it has its own standby nodes, because credit replication. And patroni can do