Re: SQL Table and RPGLE Error message

On 09-Jun-2011 07:13 , Bill Hopkins wrote:

Not seeing how trigger would cause the CPF4326 - "Commitment
definition not valid for open of."

One thing I notice about the message it does not include variable
info... i.e. file name?

So the message data is incorrect in the CPF4326; the &6 is probably a x'00', though if the error came from QDB or QDM I would expect &4 would probably have a value. That is a defect, if as issued by the IBM i OS. Perhaps the details from a spooled joblog showing that error would clarify; though perhaps only in conjunction with a WRKJOB OUTPUT(*PRINT) taken at that point in processing.

Does the output from the script with the CREATE TABLE and CREATE TRIGGER still exist? Was the script run with the default of COMMIT(*CHG)? The history log and\or joblog available from the time-frame that script was run? I infer from another defect that the message could be a side-effect of the trigger not having been created completely\correctly, such that the open of the file for update [via a logical] may incorrectly see that error message; and if, as a defect scenario, failing to set the message data properly might be a side-effect of what appears to the open processing to be a case of "resources with uncommitted changes".

If there is output generated from the request to "DMPSYSOBJ *ALL TheLibOfPF 19 D4", or if any anomaly appears for the trigger data tracked in the system database cross reference, or the RCLDBXREF *CHECK shows suspect conditions, then there is a good chance that there is something wrong with the SQL TRIGGER for the TABLE for which some recovery might be necessary.

Regards, Chuck