Cassandra at Apple: 1000s of Clusters, 300k Nodes, 100 PB

Re:Cassandra at Apple: 1000s of Clusters, 300k Nodes, 100 PB by GauntletWizard on Saturday October 08, 2022 @09:34PM (#33137578)

And nobody does reads at quorum, they're slow. People do reads at closest, even when they really need quorum.

[ Parent | on HN ]

Re:Cassandra at Apple: 1000s of Clusters, 300k Nodes, 100 PB by jjirsa on Saturday October 08, 2022 @09:53PM (#33137663)

Except they definitely do use it and many actually need it

[ Parent | on HN ]

Re:Cassandra at Apple: 1000s of Clusters, 300k Nodes, 100 PB by YZF on Sunday October 09, 2022 @12:36AM (#33138308)

I don't think that's true. Can't speak for everybody but for the stuff I worked on. Quorum reads at RF=3 are only twice as slow as reading a single node and it's pretty practical, if you care about it, to write and read using quorum writes/reads. It's true there's many applications where you're ok with some time not reading your latest data, and for those you can get somewhere better performance/latency for a given scale by going CL=ONE...

[ Parent | on HN ]

Re:Cassandra at Apple: 1000s of Clusters, 300k Nodes, 100 PB by mping on Sunday October 09, 2022 @02:48AM (#33138759)

Unless they want "read your writes", otherwise bugs start appearing "why isn't the data i just put showing"?

[ Parent | on HN ]