RecallMemoryManager is the memory manager dedicated to the recall. All recall-related files are now using RecallMemoryManager

And removed functions that were previously removed from DiskWriteTaskInterface

Changed consequently the types to use the interface instead of the derived type : RecallReportPacker

From void to bool : true if the file has been successfully written false otherwise

and thus changed the type to an auto_ptr into DiskWriteTask

RecallReportPacker is provided as a parameter of execute, and the opther attributes have been replaced by a FileToRecallStruct

Changed the constructor to take that into account, and changed the calls to the constructor where it was needed

Added first attempt to insert a real writting in a DiskWriteTask. CMakeList changed to add the File library to the tapeserver one

DiskReadTask -> DiskReadTaskInterface, DiskReadFileTask -> DiskReadTask, same thing for the write side. Changed the includes files accordingly

Changed the time interval used by the tapegatewayd daemon to know when to poll the vdqmd daemon for lost requests.  I have changed the value from 30 to 60 seconds.  30 seconds was great for testing but I beleive 60 is better for production purposes.  If we imagine each production stager has on average 60 VDQM requests pending and we have 5 production servers, then we are talking a 5Hz query rate on the vdqmd daemons which I believe is a healthy figure for a large (5 stager) CASTOR installation.

Changed the time interval between two polls on a VDQM request by the tapegatewayd daemon from 10 minutes to 30 seconds.  The tapegatewayd daemon polls a VDQM request in order to determine whether or not the request has been lost.

documented what it means (Full documentation is anyway in DynamicThreadPool.hpp).

bug #85285: tapegateway::WorkerThread::handleMigrationUpdate blindly sends a positive acknowledgements

Fixed the failing VMGR update when the tape was disabled. Reduced the cases when the vmgr helper functions refuse to work.
The vmgr helper class of the tape gateway was heavily refactored in the process.
Also added a mechanism to prevent looping on vmgr in case of a shutdown of the demon.
This was implemented with a new functor in tape/utils.
Homogenized the authorship information in tape gateway directory (to castor-dev team).

removed moved the mighunter and the rechandler in the tape directory. Deleted infoPolicy code. Made mighunter/tapegateway/rechandler to use the thread safe python library contained in castor/tape/python

The primary goal of this state machine is to help manage the complexity of the
RTCOPY (rtclientd/rtcpd) protocol within the tape aggregator.

the reception of an RTCP message into two separate function calls; the
reception of the header and then the reception of the body.

This modification was needed in order to deal with the next step of
the RTCOPY protocol where the main select loop has to deal with different
types of incoming RTCOPY messages comming from different RTCPD threads.i
Immediately after a call to select which indicates an incoming message, the
tape aggregator needs to use the Transceiever to just read in the RTCP message
header so it can then dispatch the receptionm of the body of the incomming
message to the appropriate Transceiver body reception function.

own function:

  castor::tape:aggregator::VdqmRequestHandlerThread::processJobSubmissionRequest

Started to refactorise out the giving of file lists to RTCPD.

ackowledge of a request for a VID from RTCPD.  Still a long, long way to go,
but it's a start.

tape aggregator daemon.