QVX_UNEXPECTED_END_OF_DATA

[vamsee]

Hello all, 

I am using Qlik SAP Connector 5.5.

I have a simple script to extract a particular table from SAP. 

VBAP:
LOAD 
		[VBELN] as [Sales Document_VBELN],
		[POSNR] as [Item_POSNR],
		[PSTYV] as [Item category_PSTYV]
;
SQL Select 
		VBELN
		POSNR
		PSTYV
		
From VBAP
where 
	VBELN IN ( Select VBELN from ZOPENORD)
and PSTYV IN ('ZMOS')
;

 This was running perfectly fine until a week ago, extracts around 1600 records.

Most of the time it completes in less than a minute however once in a while it takes longer.

But since last Friday I am getting an error message QVX_UNEXPECTED_END_OF_DATA.

I have read multiple posts on this error and state its due to the number of lines being fetched, whereas I am only fetching 1600.

Any suggestions on this are greatly appreciated.

@Hakan_Ronningberg  @rwunderlich @sunny_talwar @pablolabbe 

[pablolabbe]

Maybe this can help you:

https://community.qlik.com/t5/QlikView-Documents/Useful-SAP-transactions-to-troubleshoot-connector-p...

Check with SAP Admin if your transaction is being closed because it exceed some processing limit. Maybe there are too much extraction tasks running at same time against SAP system.

  I can see completed tasks with less than a minute and failed task that far exceed 20 minutes.

[vamsee]

Thanks for your suggestion @pablolabbe  Will get in touch with the SAP team to debug the issue.

Yeah, I have multiple tasks running from using the same connector, all of them run fine, apart from this one. The picture above is the task history of this one in particular. 

Attached the log you mentioned about, which does not give many details for this task though.

[Hakan_Ronningberg]

Hi,

Is there any information in the job log in SAP? Check the log for job /QTQVC/READ_DATA in transaction SM37.

Regards,
Håkan

[vamsee]

Hi @Hakan_Ronningberg  & @pablolabbe,

Apologies for the delayed response. As we use communication user and not dialog user so cannot log in to SAP Frontend to check the log SM37 for the job.

Tried and reached out to the SAP team and below is their response.

I.e. have been redirected to your reply  @Hakan_Ronningberg 

https://community.qlik.com/t5/Qlik-Connectors-Discussions/SAP-Loading-Stuck/td-p/1631890

Other options I have tried on the connection string is to reduce the packet size, increase timeout fetch, enable Log=1 and Trace=1.

Unfortunately, these did not change the result.

Again, the extract runs perfectly fine in a minute, during the morning hours but fails later on.

[Hakan_Ronningberg]

Hi,

I guess the SAP team has checked the log in SM37. It looks like the job is cancelled from QlikView.

Can you supply the SQL connector log from:
C:\ProgramData\QlikTech\Custom Data\QvSAPSqlConnector\Log 

Regards,
Håkan

[vamsee]

Hi Hakan, 

Unable to attach as the community does not support .txt files.

Below is the latest connector log after I added the Log and Trace Flags.

2020-06-25 17:12:41	Progress	Connected to SAP with C:\Program Files\Common Files\QlikTech\Custom Data\QvSAPConnector\QvSAPConnector5.5.dll 5, 5, 8698 SR3
2020-06-25 17:12:41	Error	Fetch aborted after 241 retries. Key = TIMEOUT_READ_MEMORY (ID:00 Type:E Number:001 Timeout when trying to read shared buffer)

[Hakan_Ronningberg · Accepted solution]

Hi,

Sorry for the late reply. Did you manage to fix the problem?

TIMEOUT_READ_MEMORY indicates that the background extraction job takes a long time. Can there be many other jobs in the SAP system at the same time? You can increase the time before timeout by setting a high value in the connection string property TimeOutFetch like ;TimeOutFetch=4000;

You should switch off the Trace connection string property (when it is not used) since it is bad for performance.

Also a lot of improvements has been made to the connector after release 5.5. So I recommend to upgrade to the latest release if possible!

Regards,
Håkan

[vamsee]

Hi Hakan, 

My apologies I haven't opened the community for a couple of weeks.

Yes, you guessed it right, the issue was with the SAP side. It had multiple jobs running simultaneously. I don't have many details but the SAP team has made some upgrades and it got fixed.

As suggested, we realized its high time that we update the SAP connector to the latest version.  Thanks for all your help @Hakan_Ronningberg  and @pablolabbe .

Appreciate it. 😊

[wgonzalez]

Hi,

It might be the nested Select in the Where clause.  I had a SQL Select script that included several joined tables, managing 300k + records.  I replaced the joins by loading the SQL tables into QlikView tables, then performing either joins or mapping loads in the QV script.