Cloning PDB from Single Instance to RAC results in many SYS invalid objects

Hi,

Can you post the list of invalid objects under SYS schema?

Sure, Ruchira.

This is the list of invalid objects for the SYS schema.

Hi @Manulak ,

What are the actual errors that are holding up the compilation of these errors? Can you look at a few an see whether you get more concrete errors? 

Cheers

Agreeing with Sajith (Hi Sada ;)) Similarly I don’t see these being invalid will have an impact of the objects in IFSAPP schema but SYS.CDB_INVALID_OBJECTS is bit of a concern

Do you have any errors in PDB_PLUG_IN_VIOLATIONS view?

Hello @Sajith D and @Ruchira 

Thanks for your responses.

I’m sorry - ignoring SYS invalids is never a healthy option (which is not recommended at all) which could lead to an unstable and unsecured db which has an open threat of having more and more invalids in future with cloning RAC. 

However, the question about PDB_PLUG_IN_VIOLATIONS is interesting.
Yes, there are two errors.

1. There are common XDB schema types missing from ROOT. [Resolved]
2. Database option RAC mismatch: PDB installed version NULL. CDB installed version 19.0.0.0.0.[Pending]

Probably your suggestion might puzzle out all the issues 😎 Let me check that from my side as well. But please share your suggestions aligned with the errors I’ve mentioned too.

Thanks a lot.

PS : Few more info : 4 IFSAPP objects and 3 IFSINFO objects are invalidated currently. Let’s ignore IFSINFO objects for the moment, but the errors on IFSAPP objects are clearly the invalidated referenced objects. 

The compilation errors are pretty straight forward too, i.e. the referenced objects cannot be found or not authorized. E.g. : The object IFSAPP.FND_OBJ_TRACKING_SYS has the error ORA-04023: Object SYS.MSG_PROP_T could not be validated or authorized, and so on.

Cascade compilation is recommended to be done via utlrp.sql which also hangs until it’s being killed by the hang manager. That’s why it’s suspected that this is clearly an Oracle bug which needs a standard patch, hence two SRs are being raised already.

Hello @Sajith D and @Ruchira again

About the warning in the view PDB_PLUG_IN_VIOLATIONS, our DB experts had this to say : 

Warnings in PDB_PLUG_IN_VIOLATIONS do not prevent you from actually opening the PDB in READ WRITE mode.  You can ignore WARNING messages (you cannot ignore ERROR messages).   It is okay for a PDB to have a subset (fewer) options installed than the CDB into which it is plugged.   (The reverse is NOT true, however -- the CDB must always have the same or more options as its PDBs).

Just FYI.

Your feedback is highly appreciated.

Thank you.

Hi @Manulak,

it’s interesting why you would have different features in your single instance CDB vs RAC CDB. Assuming that both were created using the IFS template, they should be the same. Was there a different route taken when the RAC DB was created?

Cheers 

Hello @Sajith D 

The installations were done as per IFS guidelines. What is the feature difference you’ve noticed? If you referred to the warning of the table “Mismatch in installed version”, can you please guide me where in IFS template that setting is mentioned \ pushed?

Thank you.

This was lately found to be an issue with the Oracle version, and certain oracle updates solved the issue.