Skip to main content
cancel
Showing results for 
Search instead for 
Did you mean: 
bnichol
Specialist
Specialist

Prevent Server Restart when Corrupt Document loaded

Is there a way to configure the Qlikview Server to prevent the Server from restarting when a corrupt document is loaded?

I have 75+ application running on a Qlikview version 9 SR4 server, and I don't want all my clients impacted if only one of the applications is experiencing problems.

Here's what was recorded in the server event log when the issue occured...

110 Warning The document failed to load because of an I/O error or a corrupt document.

131 Error DOC loading: Exception -2 while loading document

120 Error Server aborted trying to recover by restart. Reason for restart: Internal inconsistency, type D, detected.

Any suggestions would be appreciated?

B

1 Solution

Accepted Solutions
Chip_Matejowsky
Support
Support

My suggestion would be to perform diligent testing of new apps or apps elevated from lower environments prior to placing in production environments.

Best Regards

Principal Technical Support Engineer with Qlik Support
Help users find answers! Don't forget to mark a solution that worked for you!

View solution in original post

10 Replies
Not applicable

The inconsistency error can happens some time but it is hard to find why (as far as I know).

I'm not very confident in SR4 so I would, if this is possible install SR5 and test again.

If problem still happens on this specific document I would check the document Log, and server events logs.

I would also try to reload the document using Qlikview Developper.

Hope this will give you some clues

Sébastien

bnichol
Specialist
Specialist
Author

The question was...

Is there a way to configure the Qlikview Server to prevent the Server from restarting when a corrupt document is loaded?

Not applicable

The server is not supposed to restart when a document is corrupted. And As I said this error is normally rare.

This is why I would suggest to install the SR5 because I saw some strange behavior with SR4. But now I cannot guarantee you that it will solve the issue.

bnichol
Specialist
Specialist
Author

Thanks.

Gustav_Guldberg
Employee
Employee

Hi B,

If you still encounter internal inconsistencies in SR5 (reload the document with SR5 as well) please get in touch with QlikView Support so that we can help you in setting up crash logging.

Type-D is most likely related to an issue with document but it can only pin-pointed if captured properly. This can be something very simple such as a corrupt object in the layout but the best outcome is usually achieved via crash logging.


Regards,

Gustav

lhr
Employee
Employee

this is a bug. the corrupted document is not unloaded properly which in turn can cause the internal inconsistency. not loading corrupt documents is one workaround Wink

bnichol
Specialist
Specialist
Author

Is there a way to not load corrupt documents automatically? Or is there a way to automatically test for corrupted applications?

These documents are reloaded automatically, and the corruption occurs sporadically, but there is no visibility in the log?

lhr
Employee
Employee

unfortunately nothing that i know of. isn't there any errors/warnings in the distribution service log related to the tasks that ends up with a corrupted document?

fredericvillemi
Creator III
Creator III

Hello,

ten years later, I'm also interested in that topic :

I also have hundreds of applications on my server, coming from many different teams.

If one document is corrupted, because of bad upload or other reason, the server will crash because of Inconstency.  I know about the reason, probably a failed upload but what I would like is that the server would not crash when it finds a document that cannot be opened.

QlikView is really bad for mutualized servers where only one bad application will make the full server crash. Just like one bad selection on a badly created document will create issues for all users.

Is there a way to prevent that behaviour ? (don't tell me to make only perfect applications because when you have to manage dozens of developers, errors may always appear)

Thanks