Changing the Game: Moving from Reactive to Proactive High Availability

1
Changing the Game Moving from Reactive to
Proactive High Availability

• Luigi Mercone
• Senior Director, Product Strategy

2
Storage and Availability Objectives
Increase availability and DR readiness

• What isnt Mission Critical?
• Maintenance windows shrinking
• Site consolidation A-A DR

Manage storage growth

• Improve operations
• Reduce cost per GB
• Utilize existing assets

3
Agenda
4
VERITAS Cluster Server Vision
We are focused on helping you

• Drive out downtime
• Transform operations
• Leverage existing infrastructure
• Virtualize, consolidate, and scale

GOAL
5
Foundations to Drive Out Downtime
With Application, Data Infrastructure Protection
6
Driving Out Downtime with VCS
Rich ApplicationMonitoring Control
Fast Failoverwith SF Cluster File System
Non-disruptiveDR Testing
Client
Initiate Fire Drill
Mount Snapshot
Test Application
Reset
Secondary Site
Primary Site
Passive Server
Failed Server
Active Server
Active Server
Replication

• Detect failure
• Un-mount file system
• Deport disk group
• Import disk group
• Mount file system
• Start application
• Clients Reconnect

Application Service Group
BCV Snapshot
7
Agenda
8
Transform Operations Current Capabilities
CentralizedManagement
Simplified Set-Up and Training
PowerfulReporting
Roaming Spare/No SpareArchitectures for Fewer
Servers
Wizard-Driven Configuration
Clustering the HardwareVendors Way
Cluster Server Simulator
Utilization
Clustering the Symantec Way
9
Installation Assessment Service

• http//sfprep.symantec.com
• All SFHA products
• Unix / Linux only

10
Installation Assessment Service
11
Agenda
12
Leverage Existing Infrastructure Current
Capabilities
Single HA and DR Solution forLocal, Metro, and
Global Architectures
Out-of-the-BoxEnvironment Support
Metropolitan HA(Stretch Cluster)
Wide-Area DR(Global Cluster)
Local HA
Synch Replication or Mirroring
Asynchronous Replication
13
Virtualize, Consolidate ScaleCurrent
Capabilities
Roaming Spare/No SpareArchitectures for Fewer
Servers
Basic Multi-Tier withRemote Group Agent
SupportVirtualization
Server Virtualization Technology
Application Monitoring
HA
DR
VMware
Solaris LDom
Clustering the HardwareVendors Way
Solaris Zones
IBM LPars
IBM mPars
Utilization
Clustering the Symantec Way
14
Agenda
15
VERITAS Cluster Server for VMware
Hardware (x86)
Hardware (x86)
VM A
VM C
VM D
VMware ESX
VMware ESX
VCS
VCS

• End-to-End, comprehensive protection
• Deep application awareness for Exchange, SQL,
  Oracle, SAP, others
• Transparent to application administrators
• Simple to deploy and maintain
• No need for passive standby hardware and resources

16
VCS Compared With VMware HA
Cause of Failure VMware HA VCS
App processes crash/hang X
App process startup order X
Admin accidentally caused failure X
Configuration changes X
OS failure X X
NIC or network failure X
Storage connectivity failure X
Virtual Server failure X X
Physical Server failure X X
Data center failure X X
If the application fails, VMware HA is not aware
VMware HA is not aware of network or storage
connectivity failures
Future releases Experimental support (not
supported by VMware)
17
VCS One

• VCS DNA
• Service Groups
• Resources
• Agents

18
Thank You!
Luigi Merconeluigi_mercone_at_symantec.com