mcp and intel automation tools - unite• automated early notification of database issues or limits...
TRANSCRIPT
MCP and Intel Automation ToolsSteve Koss and Jim Lebak
Unisys Corporation
UNITE 2012
© 2012 Unisys Corporation. All rights reserved. 2
Agenda
• Workload Manager
• Business Continuity Accelerator
• System Assistant
• Database Operation Center
• dbaTOOLS
• Locum Real Time Monitor
• Operations Sentinel
• Opcon/XPS
MCP
• Secure Private Cloud
Intel
Workload Management
© 2012 Unisys Corporation. All rights reserved. 4
Traditional resource management
Our tried-and-true priority-based scheduling works well in many cases but3
• It requires an intimate knowledge of theapplications and their interactions
• Choosing suitable priority values can be hard
• One size does not fit all
• Change can be difficult to cope with
• What to organize workloads by business function and time period/circumstance (SLA)
© 2012 Unisys Corporation. All rights reserved. 5
Workload Management overview
Membership
rules
Workload
Workload
groups
Step 1: Membership rules are defined
to separate a workload into workload
groups
Usercode = “PROD-ON-LINE”
Example:
© 2012 Unisys Corporation. All rights reserved. 6
Example:
Workload Management overview
Membership
rules
Workload
Workload
groups Goals
Step 1: Membership rules are defined
to separate a workload into workload
groups
Step 2: Goals are applied to define
the desired performance outcome for
each workload group
Online
orders
Average response time
of .5 seconds
© 2012 Unisys Corporation. All rights reserved. 7
Workload Management overview
Membership
rules
Workload
Workload
groups Goals
Policy
Step 1: Membership rules are defined
to separate a workload into workload
groups
Step 2: Goals are applied to define
the desired performance outcome for
each workload group
Step 3: These workload groups and
their goals are collectively defined as
a policy
Daytime Policy
Example:
© 2012 Unisys Corporation. All rights reserved. 8
Workload
groups Goals
PolicyWeekend Policy
Workload
groups Goals
PolicyNightime Policy
Workload Management overview
Membership
rules
Workload
Workload
groups Goals
Policy
Step 1: Membership rules are defined
to separate a workload into workload
groups
Step 2: Goals are applied to define
the desired performance outcome for
each workload group
Step 3: These workload groups and
their goals are collectively defined as
a policy
Step 4: You can define additional
policies to accommodate differing
workload groups and goals
Daytime Policy
Example:
© 2012 Unisys Corporation. All rights reserved. 9
Goals
• Goals are the specification of the desired outcome for the workload group’s performance
• They are expressed in business terms
• There is one goal specified for each workload group specified in a policy
Workload groups Goals
Administration
Payroll
On-line orders
Default
Response time
Quota
Discretionary
Reservation
© 2012 Unisys Corporation. All rights reserved. 10
Managing to goals
• Workload Manager adjusts resourceallocations to satisfy all goals
• First choice is from available resources
• Where possible Workload Manager diverts resource from “overachievers” to programs that are not meeting their goal
• When there isn’t enough resource to satisfy all the demand, the least important goal is sacrificed
© 2012 Unisys Corporation. All rights reserved. 11
Automated Policy Events
• “Automated Policy Events” allow a Workload Management administrator to define automatic policy actions based on unexpected system events.
• The supported actions are (all combinations allowed per event):
– Activate or deactivate a Policy
– Send an e-mail
– Issue a system command
• The unexpected system events include (not a complete list):
– Hardware or Software Change
– Runaway Program Detected
– System Halt Load
– Time Zone Change
© 2012 Unisys Corporation. All rights reserved. 12
Runaway Program Protection
• User definable resource usage limits
– Resource Consumption Rate Limits
• CPU percentage and/or IO operation per second limit for a minimum time
– Resource Consumption Maximums
• Elapsed Time
• CPU Time
• IO Operations
• Database Operations
• Provides protection from rouge applications (very useful on metered system)
© 2012 Unisys Corporation. All rights reserved. 13
Host System Limits
• “Host System Limits” provides users a mechanism that allows definition of system level usage limits for CPU and I/O. Exceeding a defined limit causes a host event to be triggered alerting the Workload Management administrator.
• The usage limits can be defined at the host level using the following new “Host System Limits” dialog.
© 2012 Unisys Corporation. All rights reserved. 14
System Monitors
• The System Monitors provide an administrator a real-time visual representation of the history of changes in total system resource usage as well as the current values.
• The displays provide botha graphical and tabular representation of CPU and I/O overall system performance.
• They can be tailored to include or exclude specific data items.
© 2012 Unisys Corporation. All rights reserved. 15
Goal Monitor
• The Goal Monitor is a real-time visual representation of Workload Management’s performance against the user defined goals.
• The display provides both a graphical and tabular representationof goal compliance.
• It can be tailored to include or exclude specific data items.
© 2012 Unisys Corporation. All rights reserved. 16
Enhanced Reporting
• Reports are provided that display howwell Workload Management is doing at meeting the user specified business performance goals.
• The information is presented in a layered fashion with greater detail being provided at each subsequent level.
– System Compliance
• A single high-level system wide display of compliance
– Policy Compliance
• A per Policy summary display of compliance
– Workload Group Compliance
• A per Workload Group by Policy summary displayof compliance
© 2012 Unisys Corporation. All rights reserved. 17
Group Metered Usage Reports
• The new “Group Metered Usage” report feature provides users of metered systems reports on accumulated metered usage for selected time periods by Workload Management workload groups.
• This feature is only available from data collected that includes metering data.
© 2012 Unisys Corporation. All rights reserved. 18
Metered Usage History Reports
• The “Group Metered Usage” node has a sub-node, “Metered Usage History” that displays a graph showing the metered usage values for all workload groups over time and a tabular representation of thedata.
• The graph and the tabular data can be exported into severalformats.
© 2012 Unisys Corporation. All rights reserved. 19
Process Resource Usage Reports
• The new “Process Metered Usage” report displays individual processes by workload group showing accumulated resource usage within the selected extraction date/time range.
• The data is displayedin a simple tabular grid listing process name, mix number,CPU Minutes, MIP Months, RPMMonths, and I/OCount.
© 2012 Unisys Corporation. All rights reserved. 20
MCP Mobile Monitor – iPhone/iPad
• As part of “Application Modernization and Mobile Productivity”, the “MCP Mobile Monitor” application has been created.
• The “MCP Mobile Monitor” is an Apple iPhone/iPad/iPod application that can connect to ClearPath MCP systems and display current and historical system resource usage data.
• Data is retrieved from Workload Management and the MCP and displayed in various forms, tables, icons, pie and bar graphs, etc. on demand by the user.
Pag
e 20
Unisys Confidential © 2011/2012 Unisys
Corporation. All rights reserved.
Unisys Business Continuity Accelerator
Product Overview
© 2012 Unisys Corporation. All rights reserved. 22
Unisys Business Continuity Accelerator
• Ensures continuous availability of applications and data
• Accelerates and automates
– Relocating a workload and associated data from a primaryserver to an alternate server
– Restarting a workload on an alternate server.
• Reduces
– Downtime
– Required skill level
– Risk of human error
• Can help reduce BC costs by repurposing servers
– Example: Development/test to BC
• Makes the process repeatable and predictable
• No distance limitations
© 2012 Unisys Corporation. All rights reserved. 23
Unisys Business Continuity Accelerator
Primary Site
Image
Enabler
Operating
Environment
Business
Continuity
Accelerator Network
Alternate Site
Image
Enabler
Operating
Environment
Business
Continuity
Accelerator
Data Replication Product
© 2012 Unisys Corporation. All rights reserved. 24
Automation Avoids Manual Steps
Uses automation to
•Transfer resources
– Storage subsystem
– Software license keys
– Halt/Load parameters
– USERDATA file contents
– Network addresses
– HOSTNAME
•Activate image enabler
•Halt/Load alternate server
•Restart jobs – even if servers have different e-mode levels
Avoid manual steps - multiple utilities
•Halt both systems (Server Control)
•Make the disk drives available to the alternate server (EMC Navisphere or Control Center)
•Make partition configuration changes (LOADER)
•Change Halt/Load units and load a partition (Server Control)
•Deal with the DEAD 431 system stops
•Copy the SYSTEM/KEYSFILE (Native File Transfer, tape or CD-ROM)
•Activate Image Enablers (system console)
© 2012 Unisys Corporation. All rights reserved. 25
How fast is it?
• Unisys Business Continuity Accelerator can relocate a workload and get it running on an alternate server in as little as 15 minutes.
• The actual time for your workload will be different!
• Factors affecting recovery time include:
– Server capacity and configuration
– Number, capacity, and type of storage devices
– Number and size of databases
– Application characteristics
– Network characteristics
© 2012 Unisys Corporation. All rights reserved. 26
Business Continuity Accelerator Uses
Primary use
• Relocate a workload to an alternate server in a remote data center in the event of a disaster
Additional uses
• Move a workload to an alternate server while the primary server is unavailable due to maintenance or upgrade activities
• Migrate a workload to a new system software release while providing a secure fallback position.
• Permanently move a workload onto a newly purchased server
• Relocate a server environment to another physical server as needed for purposes such as testing, modeling, or backup
System Assistant
© 2012 Unisys Corporation. All rights reserved. 28
System Assistant
• Supervisor Program
• Highly tailorable
• Well suited for
– configuration, verification, and initialization tasks
– reacting to specific system events
– workload automation
– scheduling of recurring tasks
– ‘batching’ of operator control sequences
© 2012 Unisys Corporation. All rights reserved. 29
System Assistant
• Input File (Parameters):
– What to do
– When to do it (AT [time] or ON [event])
– Simple, block structured language
– Get compiled via the Assistant Compiler
• System Assistant:
– Run-time codefile
– Interprets the compiled parameters
• Started at system initialization, load and executes parameters
• Performs Tasks, Monitors System and Reacts to events
© 2012 Unisys Corporation. All rights reserved. 30
Example #1 - Verify configuration at initialization
ON INITIALIZATION DO
BEGIN
COMPARE
& REPORT HARDWARE;
COMPARE
& REPORT
& CORRECT SOFTWARE;
END;
System Assistant
© 2012 Unisys Corporation. All rights reserved. 31
Example #2 - Move a peripheral between systems
PROCEDURE MOVEIT;
% Operator enters “MOVEIT <peripheral to move>“
CASE OPERATORINPUT OF
“LP20”:
IF KEYIN “OL LP20” INCLUDES “INVALID UNIT” THEN BEGIN
KEYIN “ACQUIRE LP20”;
KEYIN “READY LP20”;
END;
System Assistant
© 2012 Unisys Corporation. All rights reserved. 32
Example #3 - Trap any abnormally terminated program with usercode (OPS)
PROGRAMINFO PINFO;
ON ABORT [PINFO] OF “(OPS)=“ DO
DISPLAY “Abnormal Termination: (“
& PINFO.MIXNUMBER & “) “
& PINFO.TASKNAME & “: ”
& PINFO.TASKMESSAGE;
System Assistant
© 2012 Unisys Corporation. All rights reserved. 33
DATABASE TESTDB
(NAME=MYDB ON PACK,
TOLERANCE=2);
ON NOT (TESTDB.PRIMARY) DO
BECOME_PRIMARY (TESTDB);
Example #4 – Simple RDB Takeover
System Assistant
© 2012 Unisys Corporation. All rights reserved. 34
HOST ECHOST1 (USERCODE = “SITE”);
ON BNAPA_INITIALIZATION DO
BEGIN
SEND BNA “BNAREPORTS +”;
IF BNA “NW HOSTS” EXCLUDES “ECHOST2” THEN
DISPLAY “ECHOST2 NOT NETWORKING”;
END;
Example #5 – BNAv2 Features
System Assistant
© 2012 Unisys Corporation. All rights reserved. 35
HOST ECHOST1 (USERCODE = “SITE”);
PROCEDURE TEST_COMS;
BEGIN
IF COMS “DISABLE WINDOW TEST” EXCLUDES “DISABLED” THEN
DISPLAY “TEST WINDOW DISABLE FAILED, RESPONSE = “ & COMSRESPONSE [1];
END;
Example #6 – Disable COMS Window
System Assistant
© 2012 Unisys Corporation. All rights reserved. 36
Basic Automation
Event
Recognition
Application
Recovery
System
Recovery &
ReconfigurationConfiguration
Control
System Assistant
Database Automation
© 2012 Unisys Corporation. All rights reserved. 38
Database Operations Center
• Graphical Tool for Managing a DMSII Database
• Programmatic generation of scripts to perform selected database administration operations
• The scripts can be customized and included in Workflow (WFL) jobs
– Snapshot Copies
– RDB Takeovers
© 2012 Unisys Corporation. All rights reserved. 39
dbaTOOLS Monitor and dbaTOOLS Analyzer
• Monitor and Manage DMSII Databases
• Automated Early Notification of Database Issues or Limits
– The Database administrator selects which database metrics are critical and define limits for them
– When the software detects that a limit has been exceeded, an alert can cause an e-mail to be sent or a pop-up warning on the DBA’s PC
• Automated DMSII Memory Tuning (MCP 13.1 or later)
– Database memory tuning is automated based on resource usage, system memory, and the number of databases
– The DBA can review/modify the recommended changes and decide to have them made immediately or saved for later use
Security Monitoring
© 2012 Unisys Corporation. All rights reserved. 41
Locum RealTime Monitor (MCP 13.0 or later)
• Automated monitoring of the system log for client-specified events
• When a monitored event occurs:
– Alert is raised with a client specified severity
– Client-specified action is taken (display, write to file, count, e-mail) Faster detection of security violations
• Reduce the associated business risks (costs, lost revenue, sanctions, client dissatisfaction, etc.)
• Agility to react to dynamic security events
• Improve compliance and reporting for management, regulatory agencies and auditors
Cross Platform Automation
© 2012 Unisys Corporation. All rights reserved. 43
Cross Platform Automation
• Operation Sentinel
– Automated Response to System Events
– A system administrator
• Determines which events need automated responses
• Defines the responses
– A management server
• Monitors all operator messages
• Initiates automated responses
– Responses can include
• Raising an alert
• Logging an audit trail message
• Sending a command or initiating a job (1 or more systems)
• SMA OpconXPS
– DataCenter Wide Management Tool
• Define, Schedule, Monitor, Control and Audit Jobs
Secure Private Cloud