I have hosts A, B and C. A and B are on the same LAN. C is remote.
B needs files which are both on A and C. C is running 24/7.
When I start B, it starts to synchronize over the slow remote connection from C, which is fine. Now, when I boot A, B is recognizing that A is a LAN connection but keeps synchronizing with C instead of switching to the much faster connection A.
How did you verify that?
You can probably enable model debugging to verify that, but assuming B knew that A had the full file before it came online, it should have done the right thing.
If A came online and discovered the file as being available after the download on B has already started, then this is expected.
So does a device always schedule pulling all blocks of a file from the same remote device? I thought it requested blocks from anywhere they are available, thus parts could come from A and other parts from C. Because A is quicker to deliver them, it automatically gets more requests.
The “who has the file” comes from the database, which we take a snapshot of when we start pulling. We then try to spread the requests to all devices that have the file, prioritising to devices that have the least in flight requests.
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.