Improved draining by implementing back pressure on the scheduling (no more than 1000 jobs pending per drainingJob) and ordering of files by importance (no tape and fewer copies on disk first)

        case an exception is thrown

Conflicts:

	xrootd/XrdxCastorClient.cpp

Use multiple connections to XROOT, otherwise, the server take all ur requests sequentially, and we do not manage to put any stress

        the socket in the SendAsyncRequest method

        coming back to collect the stager response after a stall
        Refactor the way the statistics are collected and drop some debug messages
        Drop useless while(0) loop - remaining from the defines
        Drop the XrdxCastor2UFS::Rem function which has the same functionality as Unlink
        Modify the DoAsyncReq method so that the user waits for 4 seconds for a response only
        the first time it submits a request. When coming back from a stall it won't wait at all
        Reduce the delay function growth

Consider replication on close as user activity so that they are scheduled in the main queue, and not in the backfill one

        we are in the DEBUG log level