1. Database Evolution and Monitoring
Liana LUPSA

2. Collaboration with Oracle
Bringing early access to Oracle database 12c Release 2 features
Evaluate infrastructure changes needed to support new features
Identify use-cases where new features can help provide better value for CERN research
In return - access to expertise and a highly demanding computing environment
Provide intensive feedback to Oracle as to help improve the quality of its products.

3. Database Evolution and Monitoring
Oracle Database 12c Release 2
Oracle Database In-Memory
Oracle Enterprise Manager 13c

4. Oracle Database 12c Release 2
New Features:
Sharding - used for applications that require extreme scalability
Refinements to Oracle Database In-Memory capabilities
In-Memory column Store on Active Data Guard
Heat Map
Support for End-to-End JSON and JavaScript
Change: Deprecation of Direct File System Placement for Oracle Cluster Registry (OCR) and Voting Files
New: Oracle Clusterware files managed by Oracle Automatic Storage Management (Oracle ASM) vs NFS
Sharding
The data is split into multiple databases (called shards). Each database holds a subset of the data. You can split the shards into multiple shards as the data volume or access to this data grows.
Shards then can be replicated for availability and scalability reasons.
Shards are replicated for availability and scalability
The ideal sharding solution should enable you to change schemas, tables, configurations, ..everything, online with high performance, painlessly -transparent for the user
-easy for backup - back-up only small databases
Starting with Oracle Grid Infrastructure 12c Release 2 (12.2), the placement of Oracle Clusterware files: the Oracle Cluster Registry (OCR), and the Voting Files, directly on a shared file system is deprecated in favor of having Oracle Clusterware files managed by Oracle Automatic Storage Management (Oracle ASM). Placing Oracle Clusterware files directly on a shared file system may be subject to desupport in future releases. If you need to use a supported shared file system, either a Network File System, or a shared cluster file system instead of native disks devices, then you must create Oracle ASM disks on the shared file systems that you plan to use for hosting Oracle Clusterware files before installing Oracle Grid Infrastructure. You can then use the Oracle ASM disks in an Oracle ASM disk group to manage Oracle Clusterware files.
If you are using NFS already, adding in the ASM layer is an unnecessary complication, but there are some situations where it might be useful:

5. Oracle Database 12c Release 2
Installation and configuration 2-instance RAC database
RAC and Security tests
Active Data Guard Multi-instance media recovery - Start media recovery on multiple instance
Data Guard DB compare
Password file synchronization to standby
Automatic Locking of Inactive User Accounts
Upgrade process to latest Oracle release from and
Role-Based Conditional Auditing
Explore new features like In-Memory on standby (read-only) databases
In-Memory on Active Data Guard
In-Memory join groups and cached expressions
Automatic Data Tiering
Active Data Guard Multi instance - Start media recovery on multiple instances, test if it behaves correctly.
Multi-Instance Redo Apply -Redo Apply can now run on all or some standby instances as configured by the user. This enables Redo Apply performance to scale, if needed, by adding additional standby instances.
Test performance improvement of running MRP on both nodes
Data Guard DB compare - detect loss writes
Role-Based Conditional Auditing - minimize audit records

6. Oracle Database In-Memory
Data held in memory in a columnar data format
Speed up analytical queries
Analytics run faster on column format
In 12cR2 - the In-Memory option will be much smarter regarding the data to hold in memory versus which data is on disk
The column store will be a cache
Repopulating managed by least recently used algorithms using the Heat Map that has been introduced in Release 12.1
A column format database stores each of the attributes about a transaction or record in a separate column structure. A column format is ideal for analytics, as it allows for faster data retrieval when only a few columns are selected but the query accesses a large portion of the data set Data in the IM column store does not reside in the traditional row format used by the Oracle Database; instead it uses a new column format. The IM column store does not replace the buffer cache, but acts as a supplement, so that data can now be stored in memory in both a row and a column format
In 12cR2, the In Memory option will be much smarter about which data to hold in memory versus which data is on disk. The column store will be a cache, with aging out and repopulating managed by least recently used algorithms using the Heat Map that has been introduced in Release In addition to this automated cache management, also policies can be defined to determine which data to keep or evict.
By using In - Memory we are able to show – sometimes dramatic –speedup of the queries
Configure which tables (partitions ) should be loaded into column store
The columns are in compressed format

7. Oracle Database In-Memory
Currently in production for the newly developed AIS data warehouse (ERP, HR & Financial data)
Testing ongoing for LHCb experiment - bookkeeping application
Join groups feature very useful
With Database 12c release 2, In-Memory also possible on Active Data Guard stand-by database
Offloading analytic reporting to standby
AWR will be available on Active Data Guard as well
Heat Map feature - the In-Memory option will be smart and able to load most frequently accessed data automatically
In­‐Memory Area becomes dynamic
Increasing size is possible without any instance restart
In-memory expressions - automatic caching of results for most frequently used function calls and expressions
By using In - Memory we are able to show – sometimes dramatical –speedup of the queries
Currently in production for the newly developed AIS DWH (ERP, HR & Financial data) - performance - coorp data stored - good for reporting
Often doing full table scans- - and the amount of data is very big - hundred of columns - after some testing the performance was improved and it went to production
By using In - Memory we are able to show – sometimes dramatic –speedup of the queries
testing ongoing for LHCb experiment - bookkeeping application
Join groups feature very useful - Data withing the in -memory is compressed - must be decompressed before join -> users specify the columns used for join when accessing tables
The app is doing queries that require the join on several tables which are big ; this feature aims at improving the performance of big joins
With Database 12c release 2, In Memory also possible on Active Data Guard stand-by database - Offloading analytic reporting to standby
Heat Map = Heat Map is a new feature in Oracle Database 12 c that automatically tracks usage information at the row and segment levels.
A new fully automated mode using heat maps decides which tables (partitions )should go into or removed from the in-memory column store
Data modification times are tracked at the row level and aggregated to the block level, and modification times, full table scan times, and index lookup times are tracked at the segment level. Heat Map gives you a detailed view of how your data is being accessed, and how access patterns are changing over time.
In-memory expressions - automatic caching of results for most frequently used function calls and expressions
Expressions have to be calculated during query execution -> 12.2 frequently evaluated expressions get added automatically

8. Oracle Enterprise Manager at CERN
Main monitoring tool for Oracle infrastructure
Improve end-users experience of CERN Applications by identifying and fixing performance issues (DB, MW)
Provide availability metrics to the CERN Service Availability status board for applications running on WebLogic servers
EM12c - current production service - Over 5500 monitored targets - over 300 hosts/agents
OEM at CERN -brief history and why do we use it ?
CERN has a long tradition of using OEM, the first version 9.2 being deployed in 2002.
To make things clear for everyone I would like to say that
Oracle Enterprise Manager is a set of web-based tools for managing the Oracle environment and it is one of the main monitoring tool used in IT-DB department for Oracle infrastructure
We use it to monitor all our Oracle estate and get instant alerts via or SMS in case of problems.
Also, OEM provides the data used to monitor and display the availability for applications servers provided by our team.
Currently we are now running EM12c R4 in our production environment.

9. Oracle Enterprise Manager 13c
Oracle Enterprise Manager 13c announced at Oracle OpenWorld - “always on” monitoring
New capabilities and features have been evaluated
Always On Monitoring Service
Gold Image Based Agent Lifecycle Management - mass deploy updates with ease
Framework and Monitoring
User Defined Target Properties
Incident Manager
Corrective Actions
Provided comprehensive feedback and reports to Oracle
Integrate with existing DB group systems and upgrading to Enterprise Manager 13c
Integration of the new Oracle Enterprise Manager 13c with the CERN Single Sign-On
We have more than 300 hosts and each host has an Oracle agent installed on it, so maintaining all this infrastructure starting from a fresh installation through upgrade and patching takes a significant part of time of an OEM administrator.
Prior to Golden Agent Images, upgrading or patching the monitoring agents took a lot of time. Now we patch and test only one reference agent and all the other agents will be patched automatically.
By using the Gold Agent feature from OEM 13c simplifies the agent lifecycle process of maintaining up to date all the monitoring agents becomes more automated and gives a high consistency.
Brownouts - monitoring of targets continues -but incidents and notifications are stopped
Always On Monitoring Service – A separate service used in planned downtime, will receive target availability and send limited notifications.
Corrective Actions – Support for corrective actions on all event types (beyond just metrics), and out-of-the box customizable Tablespace corrective action.
Target Properties Master List of Values – Define a list of values for your target properties.
Incident Manager – New dashboards and export incident rules. Enough said.
Users can now create simple definitions consisting of a reference target and comparison template which can be applied at scale and dynamically to thousands of targets. Results are reported in the new Drift Dashboard which provides a summary and detailed results of drifted targets across the entire managed estate.

10. Oracle EM 13c with CERN SSO
Fine tune the security access
What type of users can access :
CERN users
External users
SSO - forget about configuration and delegation
Redirect to the oem login page -
Single point of trusted authentication
2 factor authentication - enhanced security

11. Questions