Exadata 12c New Features RMOUG

Copyright © 2014, Oracle and/or its affiliates. All rights reserved. |
Exadata 12c Enhancements
RMOUG Training Days 2015
Fuad Arshad
North America Commercial Engineered Systems
Oracle Confidential – Internal/Restricted/Highly Restricted

• Chicagoland ,IL
• Started working on Oracle RDBMS in 1996
• Technical Reviewer, Author, Speaker, Worker Bee
Twitter @fuadar
Oracle Confidential – Internal/Restricted/Highly Restricted
Who am I

Agenda
Understanding Exadata
Exadata New Features
EM for Exadata
1
2
Oracle Confidential – Internal/Restricted/Highly Restricted 5
3

Understanding an Exadata

Oracle Exadata Database Machine Vision
• The best platform for all Oracle Database workloads
– Warehousing  OLTP  Consolidation  In-Memory Database
• Latest, most advanced hardware
– Fully scale-out servers and intelligent storage with unified InfiniBand
connectivity and PCI flash
• Unique software that maximizes the Oracle Database
– Database optimized compute, storage, and networking software
dramatically improves performance and cost
• Standardized, optimized, hardened end-to-end
– By the core Oracle RDBMS development team
Oracle Confidential - Highly Restricted 7

Redundant database
servers, storage
servers, network,
power
Active standby
systems, local HA
and/or remote DR
Log-based data
replication with data
consistency validation
Database Optimized Protection
from All Failures
–Data loss or corruption
–Unplanned outages
–Planned maintenance
–Application changes
–Human errors
Online patching,
reconfiguration,
application
changes
Active-active DB
clusters. Mirrored
storage and flash.
LAN WAN
Exadata with Oracle Maximum Availability Architecture (MAA)
Proven at Thousands of Banks, Telecoms, Web Companies, etc.
8

Exadata Components
Database
Storage Server
Infiniband Network
KVM, PDU, ILOM, CISCO SWITCH

Exadata Operational Components
DB Node i.e Linux , Database and Grid
Infrastructure Software
Storage Server i.e Storage Cell Software and
Storage
Infiniband Network i.e The Interconnect
CISCO SWITCH i.e the connection to the external
world
PDU,ILOM,KVM (X2)

Exadata 12c Enhancements

Exadata vs. Generic Hardware
Benefit Gap is Growing, and will Continue to Grow
13
ApplicationBenefits
Hardware Improvements
Common to Exadata &
Generic Platforms
• Faster Processors
• Faster, Cheaper Flash
• Larger Disk Drives
• Faster Network Interconnect
Time
• Smart Scan
• InfiniBand Scale-Out
• Database Aware PCI Flash
• Storage Indexes
• Columnar Compression
• IO Priorities
• Data Mining Offload
• In-Memory Fault Tolerance
• Direct-to-wire Protocol
• Columnar Flash Cache
• JSON and XML offload
2008 2014
• Network Resource Management
• Multitenant Aware Resource Mgmt
• Compressed Flash Cache
• Prioritized File Recovery

Copyright © 2014, Oracle and/or its affiliates. All rights reserved. | 14
 Database offload in storage
– Data intensive query operations offloaded
to storage CPUs
– 100 GB/sec SQL data throughput / rack
– Storage Index data skipping
 Database optimized PCI Flash
– Smart caching of database data
– 2.66 Million Database IOs/sec per rack
– Smart Flash log speeds transactions
 Database optimized compression
– Hybrid Columnar for 10x DB size reduction
and faster analytics
 Database optimized messaging
–SQL optimized InfiniBand protocol for high
throughput and low latency SQL
 Database optimized QoS
–I/O and CPU prioritization from application
to DB and storage
 Database optimized availability
–Fastest recovery of failed database, server,
storage or switch
–Fastest backup - incremental offload
–Exachk end-to-end validation
Exadata Innovation: Unique Software Features
Oracle Confidential – Restricted
Much Faster than Any All-Flash Storage Array + Unique DB Optimizations

Software Innovation Continues: Recent Enhancements
 Database offload in storage
– Offload searches on LOBs
– Offload joins for non-parallel queries
 Database optimized PCI Flash
– Ultra high speed flash compression
(X3 & X4) at multi-million IOs/sec
– Automatic caching for table scans
 Database optimized compression
– Hybrid Columnar enhanced for OLTP
– and for Spatial and Text data
 Database optimized messaging
– End-to-End prioritization of critical
database messages, including log writes
and RAC
 Database optimized QoS
–Prioritization of CPU and IO by multitenant
pluggable database (12c)
 Database optimized availability
–Prioritize recovery of critical DB files
–Fault Tolerant Database In-Memory
 ACFS Cluster File System
–10.2 databases, snapshots, file locks
15Oracle Confidential – Restricted

Exadata 12c
• 12.1.1.0 Software Introduced Support for Database 12.1
• Enhancements included
– Cell to Cell Data transfer
– IORM Support for Multi Tenant Databases
Next Generation Software for Next Generation Workloads

Exadata
What it looks like (12.1.x.x)

Storage Server
What it looks like (12.1.x.x)

Exadata Software: New Database Optimizations (12.1.2.1)
• Fastest OLTP
– 3x faster Direct to Wire InfiniBand Protocol
• Fastest Data Warehousing
– 5X faster Pure Columnar Flash Cache
– 3X faster JSON and XML analytics using
storage offload
• Highest Availability
– Instant server death detection
– Sub-second I/O failover caps I/O latency
– Instant failover of In-Memory Database
• Optimized Capacity
– Database Snapshots for test/dev
– Mix High Capacity and Extreme Flash
Storage
• Safest Consolidation
– Near-Zero Overhead Virtual Machines
– Per Database Flash Cache Reservation
19

OLTP: Exafusion Direct-to-Wire OLTP Protocol
• InfiniBand has great throughput
– But overheard of calling OS on every message
limits small message rate
• Exafusion re-implements RAC Cache Fusion
• Database directly calls InfiniBand hardware
– Bypasses networking software stack, interrupts,
scheduling
20
8K OLTP Block Transfers/sec
InfiniBandExafusion
InfiniBand
10GigE
3X Speedup
World’s First Database to InfiniBand Protocol

Analytics: Exadata Pure Columnar Flash Cache
• Hybrid Columnar Compression (HCC) balances
OLTP and Analytics
– 10X compression, fast analytic scans, single I/O OLTP
• Smart Flash Cache now optimizes HCC data for
pure analytic workloads
• Automatically transforms HCC data to pure
columnar format as it is loaded into flash cache
21
Hybrid Columnar Data
Flash Cache Population
World’s First Data Transforming Flash Cache
Up to 5X Analytic Query Speedup
Pure Columnar Data
Flash

Consolidation: High-Performance Virtualized Databases
• VMs provide CPU, memory, and OS isolation for consolidated
workloads
• Combine with Exadata network and I/O prioritization to
achieve unique full stack isolation
• Oracle VM on Exadata delivers raw hardware performance
– I/Os go directly to high-speed InfiniBand bypassing hypervisor
• Combine with Multitenant Database for highest
consolidation density
22
FINANCE
SUPPLY
CHAIN
SALES
World’s First Virtualized Database Platform Using Ultra-Fast InfiniBand

Exadata 12.1.2.1.0 New Features and Changes
• Database Servers
– Oracle Linux 6 (OL6.6) – Upgrade without reimage
– dbserverd service (DBMCLI)
– Virtualization with Oracle VM Server
• Stack
– Instant Server Death Detection – Lets Look rather than Wait.
– Disk Life Cycle Management Alert.
– Sparse Grid Disks and Snapshot Databases for Test/Dev
Oracle Confidential 23

DBMCLI on Database Servers
• New service dbserverd
– Replaces exadata_mon_hw_asr.pl
– MS Management Server
– RS Restart Server
– If OVM runs in dom0 only
• Command line dbmcli
• New users
– dbmadmin, dbmmonitor (welcome)
– dbmsvc (WLS owner, no login)
Oracle Confidential 24
# service dbserverd status
rsStatus: running
msStatus: running
[dbmadmin@ ~]$ dbmcli
DBMCLI: Release - Production on Tue Dec 16 01:23:48 MST 2014
Copyright (c) 2007, 2014, Oracle. All rights reserved.
DBMCLI> list dbserver
slcc12adm01 online
DBMCLI> list alerthistory

OVM on Exadata
• Support for X2-2 onwards
• Requires a Re-image , Cannot upgrade for Virtualization to be enabled
• Live Migration Not Possible due to SR-IOV
• Oracle VM 3.2.9 and OL5 ON Dom0(Hypervisor)

Patching an Exadata
• MOS 888828.1
– “The Patching Note”
– Critical Issues and additional required patches
– Guidelines for related products (OEM, DBFS, RM, GG, SAP)
• Process
– Exachk
– Patch
– Test
– Exachk

Example Cell Patching Process (Non Rolling)
• Shutdown all Databases
– crsctl stop cluster –all
– crsctl stop crs (on all nodes)
• Shutdown All Cell Services
– dcli -g cell_group -l root "cellcli -e alter cell shutdown services all“
– dcli -g cell_group -l root "cellcli -e list cell attributes msstatus,rsstatus,cellsrvstatus“
• Patch The Cells
– /patchmgr -cells ~/cell_group -patch_check_prereq -smtp_from “abc@def.com " -smtp_to “abc@def.com“
– ./patchmgr -cells ~/cell_group -patch -smtp_from “abc@def.com” -smtp_to abc@def.com
• Validate
– dcli -g ~/cell_group -l root "imageinfo“
– dcli -g ~/cell_group -l root "imagehistory“
• Bring Cluster Back up & Test

Patching Cells Prereq Check Email
Patchmgr: Non-rolling Patch_prereq Completed
Event Time 2014-08-19 09:49:08-0500
Description Non-rolling patch_prereq
completed.
Patch state of
changed from S
tarted to Succeeded.
Patchmgr launched from
Xdt104db01 s performing non-
rolling patch_prereq on following
cell(s).
7 out of 7 cell(s) completed.
Cell Patch State From Version To Version Time
Xdt104cel01 Succeeded 12.1.1.1.0.131219 12.1.1.1.1.140712 2014-08-19 09:49:07-0500
Recommended Action No action is needed.

Example Compute Patching Process (Non Rolling)
• Prepare
– Download the latest dbnodeupdate.sh from patch 16486998 & ISO from Patch 18889969
– Dcli -l root -d ~/dbs_group 'mkdir /patch'
– $ dcli -l root -g ~/dbs_group -f p18889969_121111_Linux-x86-64.zip -d /patch
– $ dcli -l root -g ~/dbs_group -f p16486998_121111_Linux-x86-64.zip -d /patch
– Validate Prereqs "./dbnodeupdate.sh -u -l p18889969_121111_Linux-x86-64.zip -v“
• Resolve
– If script mentions any nfs mount points . Dismount them. Also better to stop cluster completely as well.
– Any conflicts need to be resolved for the exact or minimum requirements also look at the obsolete list and share to make sure they are not using
it /etc/exadata/yum/obsolete.lst
• Patch
– Start patching on one node first node "./dbnodeupdate.sh -u -l p18889969_121111_Linux-x86-64.zip "
– Patch Rest of the Nodes in parallel
– dcli -g dbs_group1 -l root "./dbnodeupdate.sh -u -l p18889969_121111_Linux-x86-64.zip -q -t 12.1.1.1.1
• Validate
– Run imageinfo and imagehistory to ensure image status is “Success”
– dbnodeupdate.sh –c to cleanup and finish , Bring Cluster Back up.

EM for Exadata

Monitoring An Exadata
• Various Tools Available
• EM 12c Exadata Plugin is Recommended
• ASR Server or Platinum for Remote monitoring and Support
• Command Line Tools like Cellcli, Dbmcli, ExaWatcher , exachk

Exadata Plug-in 12.1.0.6.0: Update
Support for Oracle Database 12c
• Full support for Oracle Database
Multitenant option
• CDB level IO workload summary
with PDB level details breakdown
• DB Machine level visualization of IO
utilization for CDB & PDB
• Support for Exadata Server Version:
11.2.3.3.0, 12.1.1.1.0 ,12.1.2.1
• Support for X5 with December 14
Patches , OVM with virtualization
plugin
1
2
3

Enhanced Grid Home Page

Storage Grid Home Page
New fine grained performance summary for Flash
& Hard disk
New usage statistics highlight Flash Cache & Smart Scan
efficiency

Storage Grid Home Page
New detailed breakdown of Flash Space Usage
New IO time distribution by Flash, Hard disk & IORM Wait
per DB

Storage Cell Performance Page
New Flash & Hard Disk Utilization Chart to identify if workload
is reaching hardware limit
New side-by-side comparison of Flash & Hard Disk performance charts

Exachk and Healthcheck Utility Checks…
• Incorrect configurations
• Parameter Issues
• Recommendations, steps to correct, along with current value.
See Oracle Support Document 1070954.1: “Oracle Exadata Database
Machine Exachk or HealthCheck” and Exadata best practices in Oracle
Support Document 757552.1: “Oracle Exadata Best Practices”

Exadata OVM - exachk
• Execution model
– Run once in one dom0
• Performs database server (hardware and OS), storage server,
and InfiniBand checks
• Uses ibhosts to discover other dom0s - discovers all
database servers in logically divided racks
• -clusternodes compute1,compute2 -cells cell1,cell2,cell3
– Run once for each VM cluster
• Performs database server (OS), grid infrastructure, and
database checks

Automate Exachk via Enterprise Manager
Automating will populate healthchecks to UI!
Need to insert screenshot

Orachk Collection Manager
• Collection Manager now included in orachk MOS Note :1268927.2
Collect and Manage your System Healthchecks

AWR Warehouse – Architecture – Ideal for Exadata Data
• Central warehouse configured for long
term AWR data retention
• Historical and ongoing AWR snapshots
collected from databases enabled for
AWR warehouse
• ETL jobs moves snapshots from source
databases into AWR warehouse
efficiently using Oracle Data Pump
• Retention period configurable for
weeks, months, years or forever
(default)

AWR Warehouse Features
• Warehouse dashboard tracking ETL jobs
• All AWR features available on long term AWR
data
• Performance page
• AWR report
• ASH analytics
• Compare Period ADDM
• Compare Period Report
• Integrated seamlessly into EM UI
• Zero runtime overhead on source Production
databases

Automatic Service Request (ASR)

Exadata 12c New Features RMOUG

Recommended

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Exadata 12c New Features RMOUG (20)

Recently uploaded (20)

Exadata 12c New Features RMOUG