Datadictionary issue

After more then one year, working with Oracle support we found the issue; system is getting the DD definition from Pristine env.
Removing the Pristine OCM the issue never occurs again. It's a bug in the code (not fixed).
 
Thanks for the info. When you say "Removing the Pristince OCM" do you mean deleting all OCM records for Pristine, or just the DD Tables?
 
Just the DD, after this try to do a pkg and look at logs , you will find some errors because ocm is not found
 
All,

This issue appeared again at another install. Oracle provided a solution and it's interesting ...

As explained, the Pristine Data Dictionary is being referenced during the build. If a customer has made a change to the DD, the Pristine DD should be updated as well. They also said that this issue is fixed in 9.1.4

The original solution to remove the OCM DD mappings for the Pristine environment raised a question for me. If there is no mapping, where does it look? If this worked, then I guess we assume it's the PD dictionary? A package build is done on a path code, not an environment. What OCM is used to resolve the DD?

Interested to hear if anyone is experiencing this in a 9.1.4.x environment. My information didn't say which 9.1.4 release it was fixed in.

Craig
 
Craig,

Some idle (or not so idle) thoughts from a very, very part time CNC:

The package build is run for a pathcode, but it is still run in an environment, generally a planner environment, like DEP910. If memory serves correctly, it could get the mapping to the DD from one of three different places. 1) using the OCM for the environment in which the package build is run; or 2) using the OCM for the default environment in the JDE.INI; or 3) if the DD tables are defined as bootstrap tables (which is most likely the case), then it may get the mapping from the local database.
 
The issue is that local DD spec on ent server are updated from wrong DD (PS).
If you remove the OCM for PS they simple will not be updated in case of error (jitting from PS) ad sn you will have a good copy of local DD.

By the way pointing the OCM for PS910 env to PD910 DD datasource is better then update PS DD in my opinion.

I saw the issue in 9.1.4 too.
 
By the way pointing the OCM for PS910 env to PD910 DD datasource is better then update PS DD in my opinion

I agree that is better. Either way pristine is no longer pristine, but that is a lot less of an issue (until Oracle says "test your issue in Pristine" and the data is mess)

From the original post and others it seemed the problem occured during a full package build. That doesn't seem to be the cause. The full package will clear the local DD, opening up the possibility of JITI from the PS DD. So what process is incorectly reading the PS DD?

Oracle document 1903385.1 describes the issue and the related bug information contains an interesting line ...

The issue is happening when we JITI the specs from wrong environment, it happens randomly when there is mismatch in local environment and global environment.

Craig
 
Hi All,

I am in 9.1.4.6 and this error happens too.
Any news??

Thanks,

Eduardo
 
Back
Top