We're using QlikView 10.0.9061.7 on Microsoft Windows Server 2003 Enterprise x64 Edition X64 (Microsoft Windows NT 5.2.3790 Service Pack 2),.
I have a long-running, processor- and memory-intensive document that I set to reload last night, but it abruptly terminated partway through its script. The Document Log gave to error message (its final entry was merely a LOAD statement). The Task Log ended like this:
22/11/2012 00:07:46.1053401 Error Exception=System.IO.IOException: The semaphore timeout period has expired.
at System.IO.__Error.WinIOError(Int32 errorCode, String maybeFullPath)
at System.IO.FileStream.ReadCore(Byte buffer, Int32 offset, Int32 count)
at System.IO.FileStream.Read(Byte array, Int32 offset, Int32 count)
at QVBWrapper.Document.ReadDocumentLogThread(ILogBucket i_LogBucket, DateTime i_ReloadStartTime)
22/11/2012 00:07:46.8709945 Error Aborting Reload. Error=QDSMain.TaskResult
22/11/2012 00:11:42.7221519 Error Reload was aborted.
22/11/2012 00:11:42.7690272 Information Initializing Reload (0), Finished (15216875)
22/11/2012 00:11:42.7690272 Error Reload failed.
22/11/2012 00:11:42.8159025 Information Closing the document.
22/11/2012 00:11:42.9565284 Information Closing the document.
22/11/2012 00:11:42.9565284 Warning Document was forcibly closed. Clean close was not possible.
22/11/2012 00:11:42.9565284 Information Closed the QlikView Engine successfully. ProcessID=4584
22/11/2012 00:11:42.9565284 Information Initializing (0), Finished (188)
22/11/2012 00:11:42.9565284 Error Reload failed. Distribution skipped.
22/11/2012 00:11:42.9565284 Error The task "65 Reload of Insto Finance RB Customer Analytics\RB Customer Analytics.qvw" failed. ErrorCount=5
22/11/2012 00:11:43.2534053 Information Closed the QlikView Engine successfully. ProcessID=4584
That, however, is my only clue. The QVS logs show nothing abnormal - memory was fine in the Performance Log; no errors or memory issues in Events; there were no user sessions. There were no other tasks running.
My only guess at present is that the LOAD statement (which does an aggregate stat over the whole data set) took too long for some other process's comfort.
Anyone have any ideas over what could have caused this?