Fixed handling of d2ddest transfers put back to queue when no source remains. On the way, also fixed the case of canceled transfers where the source was also remaining

Merged into trunk fix from development branch.

bug #96388: Exception handling of the castor::vdqm::ProtocolFacade::handleProtocolVersion method is incorrect

Merged in fix from development branch.

Merged implementation from development branch into trunk.

Merged implementation from development branch into the trunk.

- Fail outstanding requests on recreateCastorFile

Fixed KeyError exception occurring after failing to start a job by a diskserver and trying to put it back to the queue

Merged into trunk the implementation from the private development branch.

Merged implementation from CASTOR2/branches/tapebridged_beyond_v2_1_13_0 into
trunk.

(and dropped internal ones as they're in the 2.1.12 branch)

other exceptions are logged at error level with their stack trace

level or above when the log is called inside an exception handler

Merged into the trunk the implementation from the private development branch /CASTOR2/branches/tapebridged_beyond_v2_1_13_0.

bug #95975: RFE: tapebridged should gracefully shutdown the rtcpd session when get next file for recall returns an error

Merged into the trunk the implementation from the private development branch /CASTOR2/branches/tapebridged_beyond_v2_1_13_0.

Merged into the trunk the implementation from private branch /CASTOR2/branches/tapebridged_beyond_v2_1_13_0.

when logging the exceptions

Made GcPolicy mandatory in the SvcClass table (not null + foreign key on GcPolicy name) + recreated SQL scripts

Merge fixed from private branch /CASTOR2/branches/tapebridged_beyond_v2_1_13_0 into the trunk.

Implementation merged in from the following private development branch:

/CASTOR2/branches/tapebridged_beyond_v2_1_13_0

plus replaced an index to optimize the draining scheduling, after
stress testing with 2M files to drain.

would be evaluated remotely by rpyc. This actually triggered an unexplained
behavior, as rpyc tried to do a remote evaluation by incorrectly calling one
of our exposed method, eventually leading to a thread waiting on the (incorrect)
nested call for the global lock it already held - and yielding to a deadlock.