High Availability Scenarios With IBM Tivoli Workload Scheduler And IBM Tivoli Framework

Skills required to implement IBM Tivoli Workload Scheduling/HACMP

There are many skills needed to place such a system into production, and it is unlikely that any one person will perform this complex task alone. A large environment requiring this type of solution generally has specialists that administer various technology sectors. Therefore, it is critical that all participants become involved early in the design process so that there are minimal delays in implementing the project.

It should be noted that while this particular exercise was specific to a IBM Tivoli Workload Scheduling/HACMP integration, the complexity and involvement needed would be no different were this a design utilizing the HP Service Guard or Sun Cluster to provide the high availability needed in a UNIX-based architecture other than AIX.

Following is a summary of the roles and skill levels needed for this effort.

Networking Administration team

The networking team must have ample time to prepare the network switches and segments for an HACMP Cluster design. They may need to supply multiple network drops at the data center floor location. Since a large HACMP configuration may require six or more network drops, there may also be a need to purchase additional switches or blades.

The skill set for these activities is medium to high. It is likely that several members of a networking team would be involved in these activities.

Required time for activity: 2 to 5 days

AIX Administration team

This team is responsible for the following tasks:

• General setup of the RS/6000 and the AIX operating systems within the cluster

• Patching AIX operating systems

• DASD configuration

• Configuring and testing of the serial heart beat cable at the OS level

• Network configuration and connectivity testing

• Possibly assisting with the HACMP and IBM Tivoli Workload Scheduling installation

The skill level for these tasks is high, and is best performed by an AIX administrator/specialist.

Required time for activity: 5 to 15+ days

HACMP Administration team

The HACMP administration team is responsible for the daily operations of the HACMP environment. Many large customers will have a team dedicated to maintaining these complicated HACMP Clusters. Some of the duties they perform are installations, upgrades, troubleshooting and tuning. It is not unusual to find them having strong AIX skills, and their duties may overlap into AIX administration.

The required skill level for these types of activities is high. The whole purpose of this environment is to provide a highly available 24-hour, 365-day a year operation. HACMP administrators having no training and a minimal skill level place the HACMP system, the application and the business at risk. Therefore, training for HACMP (or any clustering product) is required. Training for seasoned HACMP administrators is also suggested as HACMP has seen significant changes over the last several revisions.

Required time for activity: 10 to 15 days, and ongoing support

Tivoli Framework Administration team

In larger shops, there may exist a Framework team that would install the TMRs (or Managed Nodes, if you decide against a dedicated TMR) for you. This team would need to be aware that, although it is performing multiple installations of a TMR, this effort must be coordinated with the HACMP administrators.

The required skill level for this activity is medium to high. Administrators may have procedures that will make the installation more efficient.

Required time for activity: 10 to 15 days, and ongoing support

IBM Tivoli Workload Scheduling Administration team

The IBM Tivoli Workload Scheduling administration team may be well versed in the installation of the IBM Tivoli Workload Scheduling code (and patches) into the cluster. Otherwise, this task might be handled by the AIX administrators.

The skill level for this type of configuration is high. This is a process requiring a thorough understanding of the following areas:

• The IBM Tivoli Workload Scheduling application and its recommended installation procedures

• The AIX operating system

• RAID levels and file system tuning configurations

• Fundamental understanding of the HACMP environment (which introduces complexities into the normal IBM Tivoli Workload Scheduling application installation)

Required time for activity: 3 to 5 days

Hardware Purchasing Agent

This resource is responsible for purchasing all RS/6000 and AIX-related hardware, software, cables, storage cabinets, DASD, null modem serial cables, additional TCP/IP network switches and other hardware components required for the IBM Tivoli Workload Scheduling/HACMP implementation.

The skill level for this activity is estimated to be low to medium. IBM sales has resources that are capable of quickly generating a robust configuration based on a customer's general hardware requirements.

Required time for activity: 1 to 2 days

Data Center Manager

The tasks that are performed and coordinated by the data center management team can vary greatly. Tasks that need to be coordinated are floor space allocation and various procedures for placing machines into production. They also coordinate with other personnel such as electricians, HVAC specialists, and maintenance teams, who may need to prepare or reinforce the raised floor structure for the new system being delivered.

While the estimated technical skill level of this activity is low, it is an effort requiring a great deal of coordination skills. These activities can be time-consuming and need to be coordinated properly; otherwise, they will negatively impact the implementation schedule.

Required time for activity: 2 to 3 days

Electrical Engineers

Tasks performed by a licensed engineer typically deal with potentially hazardous high voltage situations.

The skill level for this type of activity is high. As this is specialized trade, it should not be performed by anyone other than a licensed engineer.

Required time for activity: 1 to 2 days

HVAC Engineers

Heating, ventilation and air conditioning configurations are generally installed in large data centers before any equipment is ever delivered onto the data center floor. As the data center equipment population grows, however, cooling requirements should be reviewed as new equipment is placed on the data center floor.

The skill level for these types of activities is high. As this is specialized trade, it should not be performed by anyone other than a licensed engineer.

Required time for activity: 1 to 3 days

Service Engineers

IBM Service Engineers (SEs) are responsible for installing and testing the base functionality of the RS/6000 and possibly the base AIX operating system. The SE may also consult with the customer and assist in such activities as:

• SSA adapter configurations and tuning

• SSA Raid configurations and tuning

• TCP/IP network configurations and tuning

The skill level for these installation activities is high. The IBM Service Engineer is a resource that is critical to a properly installed cluster configuration (for example, if a cable were improperly installed, you would inadvertently witness false HACMP takeovers).

Required time for activity: 2 to 3 days

Backup Administration team

This team provides the vital service of integrating the HACMP solution into the backup configuration. In the case of this effort, a TSM client was installed into the configuration and the cluster is backed up nightly. This team is also responsible for providing assistance with disaster recovery testing, adding one more level of security to the complete environment.

The skill level for any enterprise backup solution is high. Large backup environments require personnel who are trained and specialized in a very critical business activity.

Required time for activity: 1 to 2 days