Neoview Database Administrator's Guide (R2.2)

STOP request until they time out. You will not be able to start the same task again until at least
one of the servers has processed the STOP request and changed the task status to either STOPPED
or ABORTED.
When a data-load task is stopped abruptly, the status of the task instance is set to ABORTED.
The status of each data-load process is set to STOPPED. To determine how many records were
loaded before the task was stopped, check the log files (see “Logs” (page 115)). Data loaded before
the abrupt stop remains in the database; it is not automatically backed out.
When recovery is enabled, which is the default, the system automatically restarts the load at the
point of the failure (see “Recovery” (page 116)). If recovery is not enabled, rerunning the load
task results in the load starting again from the beginning. This results in records from the previous
load being rejected as duplicates, unless you remove the previous data from the table.
Failed Load Tasks
If a data-load process on the source system fails during a data-load task, the target table will
contain the data that was committed to the database up to the point of the failure.
If a load task fails, the task instance log reports this completion message at the end of the task
instance log:
MASTER> MASTER Execution completed with error.
The cause of the failure is indicated by error messages that were reported earlier in the log. Some
common failure scenarios are:
• You provided an incorrect name for the input file or files. The loader reports this error
message in the log:
MASTER> FATAL : RDR Error opening file file name
Check if file exists and user has permission to read it
The gelserver.log file, /opt/tandem/genus/gel/log/gelserver.log, shows an error
message indicating the cause of the failure.
• You provided an invalid format file. The loader reports this error message in the log:
MASTER> FATAL : INIT_PLUGINS Failed to initialize reader metadata.
MASTER> FATAL : MASTER Failed to initialize plugin driver(s).
• A failure occurred on the target Neoview system that caused one or more of the gelserver
processes on the Neoview system to fail. The loader reports this error message in the log.
In this example, the gelserver number five failed:
WRITER_5> FATAL : WRT_5 A fatal error occurred on the Neoview system.
Terminating the load task.
MASTER> FATAL : ThreadMonitor Thread failed
When this happens, the Linux gelserver process terminates immediately, and this causes
the other gelservers on the Neoview system that were part of this load task to time out,
waiting for data from the Linux gelserver. If this happens, you cannot resubmit the failed
load task until at least one of the Neoview gelservers has timed out and changed the task
status to ABORTED. This can take up to 30 minutes. If you attempt to rerun a failed load
task before the task status is changed to ABORTED you will see this error message reported
in the gcmd log file:
ERROR 41007: An active instance of the current task may already be
running or user doesn't have permission to start the task instance
The details of the error that occurred on the Neoview platform will be reported in the task
instance log on the Neoview platform. Contact HP Support for details.
For the first two scenarios described above, correct the error and rerun the task.
For scenarios in which the load progressed to the point where gelservers were started on the
Neoview system, this message is shown in the gcmd log file:
MASTER> INFO : MASTER Threads Started
98 Using the Neoview Loader