Native synchronization error

0
Hi All   For one of the applications I am working on (both responsive and native) we are sometimes getting synchronization errors which I cannot explain and therefore I am seeking for further advice on this forum.   Error typically sounds like this:   ll database changes executed by this action were rolled back Changing member 'A random association' of 'a random module' by user 'random user' failed for security reasons: Setting the value [MendixIdentifier:: id='random id' objectType='Entity X, entityID= 'random entity ID] for reference 'Entity X' failed, because the object for  could not be found or accessed.   After analysing the issue I can already tell that: the user for which the error is thrown has a specific modulerole within the project that has read/write access on all attributes and associations for the entities that are impacted.  (This issue is not occurring all the time and in 98% of the cases this works just perfect) The object that is mentioned in the error is found in the DB (as a matter of fact, the object is part of master data that never get's touched. It's only used for linking to operational data and thus linked in associations.   To make it more difficult, we always commit object A together with an object B. Both have an association to the same master data entity.  In the native app we this commit both objects before synching to the server. Surprisingly, object A was synched to the server without an error even though it referred to the same master data object which we also used for Object B. Object B failed and gave an error on the fact that the ID of the master data is not found or accessed. Which is quite contradictory when we have a successful synch of object A.   I looked into the JSON for both objects and they both have the same ID in the association.   Why does one synch and the other not? (I was maybe thinking of any race conditions, but I do not think this will matter. It's merely sending the object back to the server with an existing ID for 1 association)   Any help would be appreciated.  
asked
3 answers
0

Hi Kevin,

Is there any clue in entity access rights on the entities, so because of the commit of object A, the write access on object B is revoked because of some conditiion and set to only read for instance. Or your masterdata object is no longer accessible because of a read constraint?

Is your masterdata object somehow cloned&deleted somehow so its looks like the same object with same attribute values, but with a different Id?

answered
0

I'm facing the same issue. Read and Write access is in place and both entities are being committed, but issue still occurs at time. It's infrequent and I'm not able to replicate it...

 

Are there any suggestions?

answered
0

Hello,

 

Most likely, the problem is caused by timing or client cache issues with master data references. By ensuring the master data is refreshed and committing both objects in one transaction, the sync becomes reliable.

I hope this helps anyone dealing with similar issue.

 

Regards 

Reemali

answered