Scheduled reloads fail intermittently

[Anonymous]

Hello,

We have a number of scheduled reloads that run each morning around 3 - 5am. At least once a week, one or more of these fail and it's usually a similar error. I think it is something to do with the connection to the database - this is an example of a log. Has anyone else experienced this before? Is there anything we can do to stop the whole document failing? Some of them take around 2.5 hours to reload so it is becoming a big problem for users who want up to date data in the morning.

13/10/2013 03:33:32.6992540 Information 13/10/2013 03:33:32: 0269     if(index('Mach:ALAR:CLEA:GUI::DATA:ACTI:#Dat:Effe:Manu:Auto:Eagl:Modu:Spyd:',

13/10/2013 03:33:32.6992540 Information 13/10/2013 03:33:32: 0270      left(EVENT_TEXT,4)&':')=0 and IsDateFunc(left(EVENT_TEXT,10))=0,EVENT_TEXT) as EVENT_TEXT,

13/10/2013 03:33:32.6992540 Information 13/10/2013 03:33:32: 0271     

13/10/2013 03:33:32.6992540 Information 13/10/2013 03:33:32: 0272     1 as ScanRecord,

13/10/2013 03:33:32.6992540 Information 13/10/2013 03:33:32: 0273     MESSAGE_ID

13/10/2013 03:33:32.6992540 Information 13/10/2013 03:33:32: 0274   

13/10/2013 03:33:32.6992540 Information 13/10/2013 03:33:32: 0275    

13/10/2013 03:33:32.6992540 Information 13/10/2013 03:33:32: 0276   

13/10/2013 03:33:32.6992540 Information 13/10/2013 03:33:32: 0279    OLEDB CONNECT*Provider*XPassword*

13/10/2013 03:33:32.6992540 Information 13/10/2013 03:33:32:         Error: Failed to connect to QlikView Connector.

13/10/2013 03:33:33.9180888 Information 13/10/2013 03:33:32:         General Script Error

13/10/2013 03:33:33.9180888 Information 13/10/2013 03:33:33:         Execution Failed

13/10/2013 03:33:33.9180888 Information 13/10/2013 03:33:33:      Execution finished.

[Anonymous]

Sorry, forgot to mention that we are using version 11.20 SR3 64 bit.

[dvasseur]

It could be a memory/timeout issue. Do you have any memory stats ?

[Anonymous]

It could be any external issue...  If it is critical, it makes sense to set Number of tries to 2 or 3 (se task execution options on Triggers tab)

[dvasseur]

Good point, but I would use ScriptErrorMode (or something similar, I can't verify right now), It would avoid to reload the whole script

[Anonymous]

Variable ErrorMode in combination with ScriptError helps to troubleshoot.

As for the "reload the whole script " - looks like the error is at the very beginning of the script.  Just a few seconds...  In situation like this, I'd set the number of tries to 50

[Anonymous]

We definitely had a memory problem initially which caused similar errors but I have since had the ram on the server upgraded to 96GB and it never gets close to that now (peaks around 70GB) so I'm fairly confident that memory isn't the issue.

[Anonymous]

The error can be at any point in the script, usually where there is a connection to the database - sometimes it's right near the end and therefore the whole script has to be reloaded as there are a lot of resident tables etc that don't get finished off until the very last bit.

The retries options sounds good - whereabouts is the trigger tab? Is that in the QV document itself? And what's script error mode? Sorry, I'm fairly new to this!

[dvasseur]

Try something like this:

Set ErrorMode = 0;

Do

  // Only one instruction here ! Your CONNECT TO for example

  // ScriptErrorDetails is being reset after each instruction...

  CONNECT TO...

  LET vError = Replace(Replace(ScriptErrorDetails, Chr(13), ''), Chr(10), '');

Until vError = ''

Set ErrorMode = 1;

[Anonymous]

Triggers tab - it is in QlikView Management Console (QMC), go to Documents -> Source Documents, find your document and open the task.  Or, you can use "Edit This Task" (pincil icon) right from the Status tab in QMC.  There you'll see the Triggers tab.  Task Execution options is on the bottom left.