Centralize the logic of choosing worker for table into one place

[ShiKaiWi]

Describe This Problem

When carry on operations (write, flush etc.) on one table, a specific worker should be chosen. However, the logic for choosing such a worker distributes across the codebase:

https://github.com/CeresDB/ceresdb/blob/ccf988f5629436927479721b63df53b220967160/analytic_engine/src/instance/write_worker.rs#L307

https://github.com/CeresDB/ceresdb/blob/ccf988f5629436927479721b63df53b220967160/analytic_engine/src/instance/write_worker.rs#L669

https://github.com/CeresDB/ceresdb/blob/ccf988f5629436927479721b63df53b220967160/analytic_engine/src/table/data.rs#L541

Proposal

We hope the logic of choosing the worker for a specific table should be centralized in one place (one function).

Additional Context

No response

[QuintinTao]

could I ?
I want add a function in sst_util.rs
pub fn find_worker(table_id: usize, worker_num: usize) -> usize {
table_id % worker_num
}

replace the three lines you mentioned.
if find_worker(table_id, worker_num) != worker_id {

[ShiKaiWi]

could I ? I want add a function in sst_util.rs pub fn find_worker(table_id: usize, worker_num: usize) -> usize { table_id % worker_num }

replace the three lines you mentioned. if find_worker(table_id, worker_num) != worker_id {

@QuintinTao Thanks. Any contributions are welcome. The proposal is nice, and is it better to put the function find_worker in writer_worker.rs?

[QuintinTao]

could I ? I want add a function in sst_util.rs pub fn find_worker(table_id: usize, worker_num: usize) -> usize { table_id % worker_num }
replace the three lines you mentioned. if find_worker(table_id, worker_num) != worker_id {

@QuintinTao Thanks. Any contributions are welcome. The proposal is nice, and is it better to put the function find_worker in writer_worker.rs?

OK

[archerny]

issue was closed by this PR