Scenario Overview
The objective of this scenario is a subset of the core customer scenario initiative environment in which a WebSphere® Virtual Enterprise Version 6.1.1 cell is constructed using AIX Micro-Partitioning technology. The scenario focuses on the components that are directly affected by being built on top of Micro-Partitioning systems.
Audience
This document is intended for technical sales, marketing, or the interested customer who wants to learn more about some of the common features found in WebSphere Virtual Enterprise in an AIX Micro-Partitioning Environment.
Environment Details
The configuration was established to identify a series of benchmarks and points of failure within a typical environment. The environment consisted of the following installations:
- WebSphere Application Server Network Deployment Server Version 6.1.0.25
- WebSphere Virtual Enterprise Version 6.1.1 (Formerly WebSphere Extended Deployment Operations Optimization)
- Multiple custom test applications
The environment was configured with multiple applications, multiple dynamic clusters, two on demand routers, and global security enabled for selected tests.
Operational Characteristics
Implicit components used:
- Application placement controller (APC)
- Autonomic request flow manager (ARFM)
- Dynamic workload manager (DWLM)
Console operation and management:
- CPU utilization graphs
- Queuing graphs
- Application service time graphs
- Application average response time graphs
- Heap usage graphs
- Task management
- Deployment manager repository management
- Application edition management
- Server and node management
- Service policy management
- Health policy management
Configuration components:
- Global security
- High availability deployment manager
- Two physically different systems with a shared configuration repository
- Failover
- Repository checkpoint
- Application-level service policies (typical)
- A different application in each dynamic cluster
- Configured in a high complexity configuration with 7 applications
- Configured in a moderate complexity configuration with 3 applications
- Configured at the application level in which specific applications were configured and profiled for specific average response times
- Application edition manager
- Editions rolled out using different options
- Editions rolled out while under load over an extended period of time
- Conditional operation for extended period of time
- Continuous high load
- Managed operation with peak and off-peak operation
- Extended continuous operation with high load peak
- Shared Micro-Partition
- Entitled Capacity (entc): The minimum processor entitlement
- Maximum Entitled Capacity (entc_max): The maximum processor entitlement
- Shared Partition Utilization( Entc/max_entc * 100.0 ): Function that measures the total percentage of processor load of the LPAR.
Implementation Diagram
The diagram shows the maximum number of dynamic cluster configured within the cell and the relationship of the application servers in regard to the physical nodes.
The application server nodes are run on AIX LPARs.
Application Placement Diagram
The diagram shows the maximum number of applications that were configured within the cell and the relationship of these applications to the dynamic clusters and the physical nodes.
Hardware Specifications
| Machine |
OS |
# of CPUs |
CPU Speed |
CPU Type |
RAM (GB) |
Function |
Mode |
Maximum Entitled Capacity (max_entc) |
| IBM, 8844-PBN |
AIX 5.3 |
4 |
2.3 GHZ |
PPC64 |
4 |
Dmgr (Pri.) |
|
|
| IBM, 8844-PBN |
AIX 5.3 |
4 |
2.3 GHZ |
PPC64 |
4 |
Dmgr (Sec.) |
|
|
| IBM, 9117-570 (P5) |
|
16 |
1.9 GHZ |
PPC64 |
64GB |
|
|
|
| LPAR0 |
AIX 5.3 |
1 (entc) |
|
|
|
ODR |
Uncap/Shared |
14 |
| LPAR1 |
AIX 5.3 |
1 (entc) |
|
|
|
ODR |
Uncap/Shared |
14 |
| IBM, 9117-570 (P5) |
|
16 |
1.9 GHZ |
PPC64 |
64GB |
|
|
|
| LPAR0 |
AIX 5.3 |
1 (entc) |
|
|
|
Node |
Capped/Shared |
n/a |
| LPAR1 |
AIX 5.3 |
2 |
|
|
|
Node |
Dedicated |
n/a |
| LPAR2 |
AIX 6.1 |
1 (entc) |
|
|
|
Node |
Uncap/Shared |
14 |
| LPAR3 |
AIX 6.1 |
1 (entc) |
|
|
|
Node |
Uncap/Shared |
14 |
| LPAR4 |
AIX 6.1 |
1 (entc) |
|
|
|
Node |
Uncap/Shared |
14 |
| LPAR5 |
AIX 6.1 |
1 (entc) |
|
|
|
Node |
Uncap/Shared |
14 |
| IBM, 9117-MMA (P6) |
|
16 |
4.2 GHZ |
PPC64 |
64GB |
|
|
|
| LPAR0 |
AIX 5.3 |
1 (entc) |
|
|
|
Node |
Shared Pool1 |
4 |
| LPAR1 |
AIX 5.3 |
1 (entc) |
|
|
|
Node |
Shared Pool1 |
4 |
| LPAR2 |
AIX 5.3 |
1 (entc) |
|
|
|
Node |
Shared Pool1 |
4 |
| LPAR3 |
AIX 5.3 |
1 (entc) |
|
|
|
Node |
Shared Pool2 |
4 |
| LPAR4 |
AIX 6.1 |
1 (entc) |
|
|
|
Node |
Shared Pool2 |
4 |
| LPAR5 |
AIX 6.1 |
1 (entc) |
|
|
|
Node |
Shared Pool2 |
4 |
| LPAR6 |
AIX 6.1 |
2 |
|
|
|
ODR |
Dedicated-Donating |
n/a |
| LPAR7 |
AIX 6.1 |
2 |
|
|
|
ODR |
Dedicated |
n/a |
Entitled Capacity (entc): Specifies the minimum number of processors that are available.
Maximum Entitled Capacity (max_entc): Specifies the maximum number of processors available from shared pool. This statistic is available only for shared and uncapped partitions.
The Power5 and Power6 systems were constructed as separate cells for testing. Simultaneous Multi-threading (SMT) was enabled.
Verification of the Environment
Continuous load
Validation criteria
- Varied continuous load
- Service policies
- Visualization data service
- Reports
Operational profile
- Dynamic clusters and applications
Dynamic cluster name
|
Installed application
|
| DC1 |
microwebapp |
| DC2 |
microwebapp |
| DC3 |
microwebapp |
| DC4 |
microwebapp |
| DC5 |
microwebapp |
- Service policies by application
Service Policy
|
Priority |
Response time goal
|
Application name
|
| SP1 |
Highest
|
Average response 250 Milliseconds |
veApp1 SP
|
| SP2 |
Very High |
Average response 500 Milliseconds |
veApp2 SP |
SP3
|
High
|
Average response 900 Milliseconds |
veApp3 SP
|
| SP4 |
Medium |
Average response 1300 Milliseconds |
veApp4 SP
|
SP5
|
Low
|
Average response 1800 Milliseconds |
veApp5 SP
|
Validation
The verification of the environment increased load over 15 minutes and used the number of clients that were dependent on the hardware. At full load, each LPAR should reach full processor utilization. Using the lparstat command, validate the following criteria:
- The entc value that is reported by WebSphere Virtual Enterprise matches the value reported by the lparstat command. This value is valid for shared partitions only. The value is not valid for dedicated partitions.
- The entc_max value that is reported by WebSphere Virtual Enterprise is no higher than the system entc_max value. This value is valid for shared partitions only. The value is not valid for dedicated partitions.
- The entc_max value that is reported by WebSphere Virtual Enterprise converges towards the entc value as the system reaches full processor utilization.
- The shared partition utilization reported by WebSphere Virtual Enterprise is calculated as Entc/max_entc * 100 and converges towards 100% as the system reaches full procesor utilization.
- CPU Overload Protection (COP) activates based on shared partition utilization function, and appropriately queues requests.
- Service policy goals are enforced.
- DWLM weights are balanced fairly.
- All idle cycles are consumed by appropriate LPARs, up to the COP limit.All idle cycles are consumed by appropriate LPARs, up to the COP limit.
High Availability Deployment Manager Failover
Validation criteria
- High continuous load
- Visualization data service
Operational profile
- Primary deployment manager
- Secondary deployment manager
- On demand router
Validation
The kill -9 command was used to kill the specific processes in this failover scenario.
- Kill a specific process. Validate appropriate availability, failover and visual data service logging.
- Kill the primary deployment manager process:
- The secondary deployment manager takes over.
- The deployment manager is still available through the on demand router.
- The secondary deployment manager resumes visualization data service logging.
- Kill the secondary deployment manager process:
- The primary deployment manager takes over.
- The deployment manager is still available through the on demand router.
- The secondary deployment manager resumes visualization data service logging
- Kill a node agent process:
- Visualization data service logging is still available. Killed node agent statistics are not available.
- Kill the ODR process:
- Visualization data service logging is still available. Killed ODR statistics are unavailable.
- The deployment manager becomes unavailable through the ODR.
- Kill the middleware agent process:
- Visualization data service logging is still available. Killed middleware agent statistics are not available.
High Availability Deployment Manager and Application Edition Manager Failover and Recovery
Validation criteria
- Continuous load
- Application edition manager
Operational profile
- Primary deployment manager
- Secondary deployment manager
- On demand router
- Dynamic clusters and applications
| Dynamic cluster |
Applications |
| DC1 |
VersionApp 1.0, VersionApp 2.0 |
Validation
Validate that the configuration can be recovered by a standard process when the active deployment manager is killed during application edition rollout.
- Fail after synchronizing to nodes.
- Fail before synchronizing to nodes.
- Fail before the application save.
- Fail before the drainage interval of the first group.
- Fail during the drainage interval for the first group.
- Fail after the drainage interval for the first group but before starting the edition2 for first group.
- Fail during the drainage interval of the 2nd group.
- Fail after the drainage interval of the second group but before starting edition2 for group 2.
- Fail after edition2 is starting up for group2.
Follow each failure by the recovery procedure:
- Stop all versions of the application.
- Set all editions to inactive. Save and synchronize.
- Activate the previous edition (Version 1.0). Save and synchronize.
- The routing policy for the application should show that it is at Version 1.0. If the Version is not 1.0, change the version to 1.0.
- Start edition 1.0
- Send a request to the application URL through the ODR to verify availability.
Validate that the application remains available when specific processes are killed:
- Primary deployment manager goes down.
- Secondary deployment manager goes down.
- Node agent of ODR goes down.
- Node agent of application server goes down.
References
The following information is available for reference:
Application Server Network Deployment (All operating systems), Version 6.1
Application Server Network Deployment (Distributed platforms and Windows), Version 7.0
WebSphere Virtual Enterprise Version 6.1.1
Virtualization and WebSphere Virtual Enterprise white paper
