The Importance of Zero Downtime

In PGQ, three partitions. One is being used right now and another we are working on it. And third one, we use it recently and we will be processing it. This is like for from queuing in post-gas approach used by Skype in PGQ. And we could do similar thing here. But what I wanted to say also here, I would tune checkpoints. So having less frequent checkpoints would be good. Distance increases. Fewer full page writes and sorts are happening. But the price we pay here is longer restore time after crash. If people continue using it, and we are down for 15 minutes, maybe it's not acceptable. Maybe I should drop indexes before when

Transcript

Play full episode

Remember Everything You Learn from Podcasts

Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.