How to simulate a Fatal Error ?

[Former Member]

Hello MDM Experts,

On older version it was pretty easy to simulate Fatal Error by deleting some tables on repository. However using the 7.1.08 Version, this action can not be done anymore.

Do you have any tips on how to simultate Fatal Error ?

This would help our technical team to anticipate the behaviour of MDM on such circonstances.

Thanks in advance for your Help

Best Regards

Cyril S

[Former Member · Accepted solution]

Hi Cyril,

I guess you feel this way because MDM had lesser features then whereas they have been hugely enhanced.

To do a Root Cause Analysis(RCA) you have to have the basis of a Fatal error in hand.

This one can get from Logs/Reports/Traces/Core files etc.

Once you have done this you know exact reason.

Post this you can go backwards with the series of events and reproduce the error.

Important thing is the RCA and elimination of other possibilities.

Also,since it is fast evolving tool 100% anticipation is diificult but as I said RCA and SAP notes can serve you good.

Hope this helps.

Thanks,

Ravi

[Former Member · Accepted solution]

Thanks Ravi for your quick answer.

The thing is i would like to create an error by myself. So I know that before the deletion of a Reference table was consider as an error (not sure but probably a fatal one).

It 's now pretty difficult to do so, i have tried to import a schema. I have accepted the Field creation but not the linked table.

But on new version, some controls are done before the import....

Do you see my point ??

Best Regards

Cyril S

[Former Member · Accepted solution]

I just want the system to crash with Fatal and Non Fatal Errors.

How can I do so ??

I know that we are suppose to do the inverse, but it's a way to anticipate and see if we can trace those failures.

Cyril S

[Former Member · Accepted solution]

Hi Cyril,

Thanks for the details,I see your point.

Suppose you talk of an incident which causes the system to abort operation when trying to import schema.

One has to look into MDS logs\traces and see if what details are present for that error.

If one gets that find out why did the system abort,one reason could be MDS overload.

Then you can decrease the load by removing a few component maps from the schema and retry.

If it works,You have 2 cases,one working and one not working and you also have RCA.

The point I am trying to make here is its difficult to find exact issues,looking Logs/reports/traces give you hint.

And you can get major help from MDM-BASIS area and tools.

Tools like Wiley Introscope can help monitor the performance of system.Also it would be a healthy practice to take a dipstick of load and system behavior on timely basis.

In case of few Fatal errors you just have to wait for system to crash and generate core file for analysis.

Logs dont tell you the picture there.

For Non Fatal errors a Repair certainly helps and reports help you to zero down on the errors and warnings present.

I totally agree with you,one can anticipate most of the issues and be prepared for such issues in future.

Please revert in case of a query.

Thanks,

Ravi