Closed ICGog closed 8 years ago
CL 266326 addresses this, but we're planning on moving to a unified Quincy cost model that abstracts the distributed file system through a universal interface that works for the simulated GFS (in simulation) and real-world HDFS (when running on real clusters).
I'll leave the issue open until this is implemented, but change the title to a more accurate one.
I believe this is fixed as of CL 266533 (0b8305f3e25c02d0a87349a86e4678bc06f20344). Good to close?
Yes, we've merged them.
The simulated Quincy should wrap around the real Quincy model and feed it with simulated data locality information.