Guidance to interpreting qmc logfile failures - Qlikview 11.2

[timsaddler]

Hello all

I have a job that fails when auto running in the QMC , but always succeeds when re-run from the console.

What is the best way of trying to fault find the issue ?

How do I translate  Apply rules failed. ---> System.InvalidOperationException: Operation is not valid due to the current state of the object

I enclose the logfile section below.

If anyone could point me in the right direction , I'd be grateful.

Thanks

(2017-02-15 04:20:13) Error: The task "Reload of Production Accesspoint\AI BatchPanel Interface with Barcus" failed. Exception: || QDSMain.Exceptions.TaskFailedException: Task execution failed with errors to follow. ---> QDSMain.Exceptions.ApplyRulesFailedException: Apply rules failed. ---> System.InvalidOperationException: Operation is not valid due to the current state of the object. || at QDSMain.LoopReductionSet.get_Current() || at QDSMain.DistributeTask.ClearSelections(Document qvbDocument) || at QDSMain.DistributeTask.ApplyRules(Document qvbDocument, Int32& round) || --- End of inner exception stack trace --- || at QDSMain.DistributeTask.ApplyRules(Document qvbDocument, Int32& round) || at QDSMain.DistributeTask.DistributionExecution(String fileName, Document qvbDocument, DistributionRequest distributionRequest, TaskResult taskResult, String repeatVariableName, String currentRepeatVariableValue) || at QDSMain.DistributeTask.Distribution(String fileName, DistributionRequest distributionRequest, TaskResult taskResult, String repeatVariableName, String currentRepeatVariableValue) || at QDSMain.DistributeTask.Execute(TaskResult currentTaskResult) || --- End of inner exception stack trace --- || at QDSMain.DistributeTask.Execute(TaskResult currentTaskResult) || at QDSMain.Task.AbstractTask.TaskExecution(ILogBucket logBucket, TaskResult taskResult)

(2017-02-15 04:20:13) Information: Task Execute Duration=00:08:54.7870281

(2017-02-15 04:20:13) Information: Sending Alert mail

[michael123]

Split the task into one reload and one distribute task to see where the error occurs. If it is in the distribution part I would try without any looping of fields or sending of parameters and adding them one by one to find out what excactly is causing the error.

Regards,

/Michael

[Chip_Matejowsky]

Hi Tim,

In the log file snippet you provided, what occurs above the error listed? Does the source document reload fail or succeed? If the source document itself fails, then enable document logging on the QVW so that a document/script log is generated. Also, what type of source data is the reload pulling from? If from a backend database, is there a connector involved? If so, then there could be a connector log as well to provide more information.

Are you opposed to posting the entire task log? If not, posting the whole log might be more informative for those of us who attempting to assist you.

Best,

-Chip

[timsaddler]

Yes I can post the whole log if necessary.

The job has run without failure today.

Whether this is due to a cluster reboot undertaken last night , I don't know, but I have noticed in the past that performance of reloads and distributions improves after a reboot.

I will monitor the situation....

[Chip_Matejowsky · Accepted solution]

Ah, thanks for providing response. Is this a single node QlikView installation, where all QV services are running on one server? If the reload/distribution is successful after a reboot or service restart, and then begins to fail after a period of time, you might be running into a system resource issue.

If this is a single node installation, I would check your working set limits for QlikView Server and depending on how much system RAM is installed on the server, lower the Lower Working set downward to free up some RAM from the QlikView Server Service (QVS) so that the QlikView Distribution Service (QDS) is able to perform. Provided the server has enough RAM, this would be a short-term work around. A long-term resolution would be to move the QDS and QlikView Management Service (QMS) to their own separate server, so that they have ample resources to function as they should.

Hope this helps

-Chip

[timsaddler]

Hi Chip,

Thanks for the prompt response. I am looking at creating a separate server at some point.

currently we have 2 physicals in a cluster both with 256 gig RAM.

Is there a doc somewhere that details how to do this?

Also the "reload & distribution" server would be a virtual.

is it ok to mix physicals and virtuals?

cheers