FranMorilloAWS
FranMorilloAWS
Can you explain more on that process, so based on the interval of my delta commits, Flink will run a full checkpoint?
But also wondering if my application is meant to be stateless, why is checkpoint size increase tenfold? Is that the index size? how can i manage this
But could you share what is kept in state in a Flink Hudi Application? I have just a source generator within flink and writing into hudi using MOR and state...
Is there any state cleanup that can be done? So it doesnt indefinitely ?
Then how would you recommend to work with Hudi Tables that can grow to infinite rows?
Are there any improvement or features coming for bucket index, to allow updating multiple partitions or changing the number of buckets?
Does this allows to be updating multiple partitions or if the partition changes? Is it a global index?
How is the bucket number automatically expanded? I saw in the pfr that it is meant to be a subtask in the clustering service, but clustering only works with COW...
Do you have an example? Would this only work for adding a new field, or could it work as well by modifying a column type?
And also modify the schema in rowdata or capture the schema from a schema registry and then use it in the sink? Wondering if there is any other writer that...