5.31.2018

Oracle 18c Autonomous Health Framework (AHF) - Part 1

Recently I had to present at eProseed AnualTech conference in Luxembourg and I was requested  to present a topic something about Oracle18c.

Obviously I don't want to talk and repeat the same about Autonomous Database, many experts already said much on this. I then decided to pick a topic which really helps DBAs, Administrators and finally to the organization. I was really fascinated about Oracle 18c autonomous health framework concepts and decided to do a presentation on this topics.

Working in a complex and huge Oracle environment, I knew where most of our energy and time is spend, as a DBA or system administrator. We always focus on avoiding run time availability and performance issues. In a complex and critical environment, every other day, you will face a new challenge and you must be on your toes as DBA during the business hours.

For DBA, most importantly, we need to ensure the database availability, at the same time, ensure its deliver the same performance 24x7. Imagine, if you get stuck with latches, instance crash, node crash, someone changes the binaries permission/ownership, you will spend hours and hours to fix and find the root cause of the issues.

With 18c autonomous health framework, its easy to avoid and auto fix run-time availability and performance issues. There are 8 components that makes this framework. Though some of them are present in 12.2, but, all these are configured automatically upon 18c configuration and run 24x7 in daemon mode. Also, 3 of the components have machine learning capabilities. to fix issue automatically.

I will start discussing about 8 components in next series. Stay tune for more on this topic.



5.30.2018

utlrp weird behavior with INVAID objects in an Orace EBS database

In one of the recent database migration & upgrade activities, I have migrated an Oracle EBS database 11.2.0.4 to a new DB host and upgraded to 12.1.0.2. After migration and post upgrade, an utlrp.sql was ran to validate the 100k invalid objects in the database.

Weirdly, during the utlrp execution and when the INVALID objects count goes to 200, the INVALID objects counts started to increase again, and ultimately reaching the 100k number again. The utlrp was in kind of loop and never exited.

As a workaround, I have manually compiled all invalid objects and ran ultrp which ran successfully. I am yet to unlock the theory caused the situation.

Stay tuned for more updates on this post.

Traditional Export/Import Fail With ORA-01031 in DB Vault Environment Starting Release 11.2.0.3

Oracle Database Vault was configured on RAC database v11.2.0.3 and appropriate permissions given to allow the regular duties, such as data pump export/import etc.

However, when classical/traditional import (a dump file from v9 database) was performed, the import on 11.2.0.3 database failed with the following errors:


Import: Release 11.2.0.3.0 - Production on Tue May 29 16:56:23 2018

Copyright (c) 1982, 2011, Oracle and/or its affiliates.  All rights reserved.


IMP-00058: ORACLE error 1031 encountered
ORA-01031: insufficient privilegesUsername: 

Data pump jobs were successfully executing without any issues. The only problem is with traditional export/import . According to below Oracle notes, it is an expected behavior in 11.2.0.3 with DB Vault in place.


Classic Export/Import Fail With ORA-01031 in Database Vault Environment Starting Release 11.2.0.3 (Doc ID 1419687.1)
Bug 13615338 - EXP/IMP disabled in Database Vault environment (Doc ID 13615338.8)

According to the notes,

Export and Import tools are no longer recommended to be used, especially in Database Vault environments.
When it is not possible to use data pump (for instance when the export has been created from a pre-10g database),
the patch 13615338 must be installed to allow the usage of classic export/import utilities.



The option would be to upgrade the database to higher than 11.2.0.3.


Will apply the patch and update this blog if the issue is fixed or not.