This involved moving the definition of basic exceptions to the libcastorcommon and thus adding dependencies to this library in a couple of places that depended only on libcastorclient

Removed the code and structures not used anymore as a consequence of the move to bulk.

Finished the refactoring of the code the bulk calls to the DB in tape gateway. As a result, there is a lot of dead wood in the code, still to remove. But the testing can start now.

Forgotten modifications from test environment.

Finished the refactoring of the code the bulk calls to the DB in tape gateway. As a result, there is a lot of dead wood in the code, still to remove. But the testing can start now.

Bugfix for the retry policy. Now it works also if we have only a
standard unsecured nsd listening.

Implementation of the new policy of trying to connect securely to nsd first and then falling back to the normal, unsecured connection if the former fails.
This should be completely transparent for the user.

Rewrite the db interface and worker thread of tape gateway to implement the bulk update inside the DB.
Interim check-in. Bulk is still a work in progress.

DB interface relies on setVector in OCCI, which requires the OBJECT option to be enabled in the client.

Also:
Fixed remaining class declaration for Tape.
Fixed modifications of messages in DLF which.
Expanded the occiNumber utility class to bring full 64 bits support without the limitations of the cast to a double (was 51 bits).

Done again, but this time there is an upper limit for the number of retries for
all types of error.

packages that need Globus.

Done.

Fixed packaging for the security plugin: now the castor-csec RPM is gone
and the KRB5 plugin is part of the castor-lib RPM.

Defaults have been modified so that the buggy GSI plugin doesn't get compiled unless explicitly re-enabled

Fixed sr #128448: Hanging transfers in listtransfers (c2cms/t0temp). Basically the problem was due to disk to disk copies that could not find any destination

  - many changes in the DB schema :
     + Tape and Segments are gone
     + RecallJobs have drastically changed
     + RecallMount, RecallUser and RecallGroup have appeared
     + SubRequest's use of WAITTAPERECALL has changed
     + DiskCopies's status WAITTAPERECALL has disappeared
  - drop of all recall python policies (and the associated thread in TapeGateway)
     + replaced by PL/SQL code
  - drop of several generated C++ objects : Tape, Segment, RecallJob
  - modification of others, in particular CastorFile

Fixed bad error message in diskmanagerd leading to an exception and the left over of some tranfers in listtransfer output.

Added #include <unistd.h> so that the code can compile on the following version of gcc and use the usleep() function.

gcc version 4.7.0 20120414 (prerelease) (GCC)

Fixed.

Changed char* into std::string to make SLC6 happy

Dropped unused code

bug #94256: RFE: Add rule for bad checksums to the example retry policy file /etc/castor/policies/migrationRetry.py

Improved the fix.

bug #94256: RFE: Add rule for bad checksums to the example retry policy file /etc/castor/policies/migrationRetry.py

Fixed.

bug #92460: tapebridged should gracefully shutdown a migration tape-session when tapegatewayd reports a disabled tape

Fixed.

Fixed startMigrationMounts procedure that was looping forever in the present state due to bad use of counters

query for migration. Previous version was poorly behaving in CMS
as analyzed with Nilo.

Moved all of the DLF parameter type 'defines' into the same file (h/dlf_api.h) and fixed freeing of a parameter name for raw parameters that don't have a name.

It works with the caveats that the OBJECT bit should be set in the OCCI environment and the types conversion are defined at the global level, not in a package.

Fixed a silly file-descriptor leak in the following unit-test that I used to
find the memory leak of the bug-report:

castor::tape::tapebridge::ClientProxy::testSend1000FilesToMigrateListRequestMessages()

Please note that the file-descriptor leak was a bug in the unit-test and not in
any production code.

Fixed.

I have also added the following unittest:

castor::tape::tapebridge::ClientProxyTest::testSend1000FilesToMigrateListRequestMessages()

This test as its name suggests sends and receives a
tapegateway::FilesToMigrateListRequestMessages message 1000 times in order to
facilitate looking for memory leaks.

Please note that the memory leak fixed here was per converter created per
thread and that there was no memory leak per marhsalling or un-marshalling of
the tapegateway::FilesToMigrateListRequestMessages message.