- 12 Mar, 2009 13 commits
-
-
Steven Murray authored
-
Steven Murray authored
will be necessary when retrieving a disk file from an aggregation. The logic behind this being each disk file has its own unique position and length within an aggregation on tape and therefore, as far as recalling a disk file is concerned, there is no concept of aggregation.
-
Steven Murray authored
-
Akos Frohner authored
-
Steven Murray authored
-
Steven Murray authored
Renamed TapeDiskRqstHandlerThread to TapeDiskRqstHandler because it is not a thread, it is in fact a functor that the TCPListenerThreadPool uses to define the task of its threads
-
Nicola Bessone authored
-
Steven Murray authored
-
Steven Murray authored
-
Dennis Waldron authored
-
Dennis Waldron authored
-
Dennis Waldron authored
-
Dennis Waldron authored
-
- 11 Mar, 2009 3 commits
-
-
Steven Murray authored
-
Giuseppe Lo Presti authored
-
Steven Murray authored
-
- 10 Mar, 2009 10 commits
-
-
Sebastien Ponce authored
-
Sebastien Ponce authored
-
Sebastien Ponce authored
-
Sebastien Ponce authored
-
Steven Murray authored
-
Steven Murray authored
-
Steven Murray authored
-
Dennis Waldron authored
-
Steven Murray authored
-
Dennis Waldron authored
Fixed a bug in the startRepackMigration procedure which incorrectly set the size of a diskcopy. Prior to this fix the size was being set to the value of the filesystem id
-
- 09 Mar, 2009 8 commits
-
-
Steven Murray authored
-
Steven Murray authored
and function strings to tape exceptions. Example usage: TAPE_THROW_EX(castor::exception::Internal, ": Failed to down cast reply object to tapegateway::FileToRecall"); TAPE_THROW_CODE(EBADMSG, ": Unknown reply type " ": Reply type = " << reply->type()); Fixed bug in RTCOPY protocol; we were sending one too many "end of file lists". Fixed bug in message dispatch logic, there was a missing break in one of the switch statements that was causing an IObject to get dynamic casted to two different classes.
-
Giuseppe Lo Presti authored
opened in all cases.
-
Giulia Taurelli authored
-
Giulia Taurelli authored
-
Giulia Taurelli authored
-
Steven Murray authored
-
Nicola Bessone authored
-
- 08 Mar, 2009 2 commits
-
-
Dennis Waldron authored
-
Dennis Waldron authored
-
- 06 Mar, 2009 4 commits
-
-
Giuseppe Lo Presti authored
-
Sebastien Ponce authored
More reasonnable default timeout (2 days instead of 30). Note that using 30 days was overflowing the unsigned integer taken by the poll system call for the timeout. Indeed, the timeout is given in miliseconds and thus the maximum value is around 28.4 days :-) With 30 days, the number went negative, meaning (under linux) we were having a blocking call.
-
Dennis Waldron authored
-
Dennis Waldron authored
-