In large-scale replication work, there often is a problem where all permitted downloads are taken up by a slow server with big files. But to get a good transfer rate, you must keep all data nodes active at all times. The simple way to do this is to set a maximum number of parallel downloads separately for each data node. This does not preclude setting an overall maximum as before.
In large-scale replication work, there often is a problem where all permitted downloads are taken up by a slow server with big files. But to get a good transfer rate, you must keep all data nodes active at all times. The simple way to do this is to set a maximum number of parallel downloads separately for each data node. This does not preclude setting an overall maximum as before.