multitenant

Unplug an Encrypted PDB (ORA-46680: master keys of the container database must be exported)

In the Oracle Database Cloud DBaaS you provision a multitenant database where tablespaces are encrypted. This means that when you unplug/plug the pluggable databases, you also need to export /import the encryption keys. You cannot just copy the wallet because the wallet contains all CDB keys. Usually, you can be guided by the error messages, but this one needs a little explanation and an example.

Here I’ll unplug PDB6 from CDB1 and plug it into CDB2

CDB Views and Query Optimizer Cardinality Estimations

Today I faced a performance problem caused by a bad cardinality estimation involving a CDB view in a 12.1.0.2 multitenant environment. While solving the problem I did a number of observations that I try to summarize in this blog post.

First of all, when checking the execution plan of a query already running for more than two hours, I noticed that, in the execution plan, neither the referenced CDB view nor one of its underlying objects were referenced. The following query (and its execution plan) executed while connect to the CDB illustrates (I also added the 12.2.0.1 output to show you the difference it that area):

12c Multitenant internals: PDB_PLUG_IN_VIOLATIONS

In the previous post https://blog.dbi-services.com/12c-multitenant-internals-pdb-replay-ddl-for-common-users I’ve done some DDL on a common user to show how this is replayed later for PDBs that were not opened at that time. But what happens when one of the DDL fails on one PDB?

12c Multitenant internals: PDB replay DDL for common users

In multitenant, you can create common Users, Roles, and Profiles. You create them in CDB$ROOT, with the CONTAINER=ALL clause (which is optional because it is the only possible value when connected to CDB$ROOT) but they are visible to all containers. As the goal of multitenant is to avoid to duplicate common metadata to all containers, You may think that they are visible through those magic metadata links. But that’s actually wrong: they are simply replicated with a very simple mechanism: the DDL for common objects is replayed into each user PDB.

I’m connected to CDB2’s CDB$ROOT and I have two pluggable databases:

SQL> show pdbs
 
CON_ID CON_NAME OPEN MODE RESTRICTED
------ -------- ---- ---- ----------
2 PDB$SEED READ ONLY NO
3 PDB1 READ WRITE NO
4 PDB2 MOUNTED

PDB1 is opened and PDB2 is closed.

Drop PDB including datafiles may keep files open

I like that Oracle drops the datafiles from the operating system when we remove them from the database (with drop tablespace or drop pluggable database) because I don’t like to have orphean files remaining in the filesystem. However, to ensure that space is reclaimed, we must be sure that Oracle did not leave a process with this file opened. Linux allows to drop an open file but then drops only the inode. The consequence is that we do not see the file, but space is not reclaimable until the process closes the handle.
Here is a case where I’ve had an issue in 12.2 where plugging a PDB is done in parallel and the parallel processes keep the files opened even if we drop the pluggable database.

I have 1.2 GB free on my filesystem:
SQL> host df -h /u02/oradata
Filesystem Size Used Avail Use% Mounted on
/dev/mapper/ol-root 3.1G 1.9G 1.2G 61% /

12c Multitenant Internals: compiling system package from PDB

DPKi1vxX0AAADLmWhen I explain the multitenant internals, I show that all metadata about system procedures and packages are stored only in CDB$ROOT and are accessed from the PDBs through metadata links. I take an example with DBMS_SYSTEM that has nothing in SOURCE$ of the PDB. But I show that we can compile it from the PDB. This is my way to prove that the session can access the system objects, internally switching the session to the root container when it needs to read SOURCE$. At DOAG Conference I had a very interesting question about what happens exactly in CDB$ROOT: Is the session really executing all the DML on the internal tables storing the compiled code of the procedure?

Multitenant internals: INT$ and INT$INT$ views

This month, I’ll talk – with lot of demos – about multitenant internals at DOAG conference. CaptureMultitenantInternals

Multitenant dictionary: what is consolidated and what is not

The documentation says that for Reduction of duplication and Ease of database upgrade the Oracle-supplied objects such as data dictionary table definitions and PL/SQL packages are represented only in the root.

Unfortunately, this is only partly true. System PL/SQL packages are only in root but system table definition are replicated into all PDBs.

This post is an extension of a previous blog post which was on 12cR1. This one is on 12cR2.

Got published in the Red Stack Magazine

The November 2017 issue of the Red Stack Magazine published by DOAG and AOUG contains one of my articles about Oracle 12c Multitenant </p />
</p></div>

    	  	<div class=

Wrong result with multitenant, dba_contraints and current_schema

Multitenant architecture is not such a big change and this is why I recommend it when you start a project in 12c or if you upgrade to 12.2 – of course after thoroughly testing your application. However, there is a point where you may encounter problems on dictionary queries, because it is really a big change internally. The dictionary separation has several side effects. You should test carefully the queries you do on the dictionary views to get metadata. Here is an example of a bug I recently encountered.

This happened with a combination of things you should not do very often, and not in a critical use case: query dictionary for constraints owned by your current schema, when different than the user you connect with.