- 02 May, 2014 2 commits
- 30 Apr, 2014 9 commits
-
-
Eric Cano authored
-
Eric Cano authored
Added missing call to finish for the memory manager in MigrationTaskInjector. Added missing setting of task injector to the disk read thread pool in MountSession::executeWrite. Fixed missing setting of m_vid member variable in TapeWriteSingleThread. Fixed settig of wrong parameteres in castorConf variable in MountSessionGooddayMigration unit test. Added validation of the migrated files checksums in MountSessionGooddayMigration.
-
Eric Cano authored
Otherwise, there was a race condition between the control thread and the worker threads. Now, the log context of the worker thread is passed and used by popAndRequestMore.
-
Eric Cano authored
in ClientProxy.
-
David COME authored
-
Eric Cano authored
-
Eric Cano authored
-
Eric Cano authored
This unit test still needs additional checks for completion (currently commented out).
-
Eric Cano authored
-
- 29 Apr, 2014 5 commits
-
-
David COME authored
-
David COME authored
the bytes/files numbers are specified at construction time and thus no longer necessary at the call
-
Eric Cano authored
The session mode (and client type) is now construction parameter. We now support migration requests as well. A new facility records the checkums received in migration notifications.
-
David COME authored
Got rid of BlockingQueue::tryPop and changed the memory managers destructor to use instead popGetSize
-
David COME authored
Added a few logs in DiskReadThreadPool ans fixed some log levels into MemManager MigrationTaskInjector and RecallMemoryManager
-
- 28 Apr, 2014 13 commits
-
-
David COME authored
-
David COME authored
-
https://git.cern.ch/kerberos/CASTORDavid COME authored
Conflicts: castor/tape/tapeserver/daemon/MountSession.cpp castor/tape/tapeserver/daemon/TapeWriteSingleThread.hpp
-
David COME authored
-
David COME authored
-
Eric Cano authored
-
David COME authored
-
David COME authored
-
David COME authored
In case of global error during migration, all requests (old or new one) for jobs are discarded
-
Eric Cano authored
when failing to signal the end of an empty session.
-
Eric Cano authored
It now inherits from the simpler TapeSingleThreadInterface<TapeWriteTaskInterface> instead of TapeSingleThreadInterface<TapeWriteTask>.
-
Eric Cano authored
-
Eric Cano authored
It will allow flushing of potential leftover requests at the end of as session and a controlled thread end.
-
- 27 Apr, 2014 1 commit
-
-
Eric Cano authored
-
- 25 Apr, 2014 5 commits
-
-
David COME authored
in order to no even move the little finger if a previous task failed Also added some comments
-
David COME authored
-
David COME authored
Move to a static member function of payload and renamed it zeroAdler32
-
David COME authored
-
David COME authored
If a Diskread ou TapeWrite task fails, all others stop and clean memory blocks while the task injector discards all remaining requests
-
- 24 Apr, 2014 5 commits
-
-
David COME authored
-
David COME authored
-
David COME authored
-
David COME authored
-
David COME authored
-