Fixed bug #105006: Draining disk 2 disk copies are not always repecting the GuaranteedUserSlotsPercentage

Fixed bug #104998: Synchronization of d2d srcs too aggressive and creating issues for draining cancelation

        backup exists flag displayed when doing the stat through the XRootD interface. For
        backwards compatibility the Offline flag will still be set but this one has a different
        meaning in the XRootD world, namely 'file is in cache'.

        applications (e.g. ROOT) can add to the user opaque information.
        Reword some log messages.

Conflicts:
	castor/db/oracleStager.sql

        for the authorization component (acc). Use the xcastor_* logging only after
        the logging infrastructure has been properly enabled.

function and build the XRootD path inside the rtcpd_xroot_stat function. This will also append
the opaque information used for the authorization check at the XRootD diskserver.

        masked/unimplemented file system operations

ReadFile::Position is now split in 2 smaller functions and the read session is corrupted (when positioning by Feq) if something wrong happens in the ReadFile's constructor or if we are in a bad state in the destructor

Fixed case where a gridFTP client sends overlapping chunks. We now raise an error instead of screwing the checksum

Restaured ZMQPollEventHandler's virtual destructor and commented out the cleaning of the reactor in the fork process (which is dead code once the forker will be used for these sessions)