Any recommendation about data placement solver parameters?

[r-value]

Given node number N in cluster, is there any recommendation about placement solver parameters like replica factor?

I tried to solve data placement for 127 node and 5 replica and the solver yields a huge model with about 2e7 rows, 5e6 cols and 5e7 non-zeroes. The default HiGHS solver just timeouts after hours of computation.