DB2 Health Checks: How to Keep Your Mainframe in Good Condition
In order to make sure that mainframe systems function to their full potential, it is essential not only to carry out daily maintenance but also to perform regular checks that allow dangerous inefficiencies to be detected. "One such solution is the health check service, which involves regular system audits to eliminate threats posed by non-optimal configuration settings or changing software versions (including databases) . It is not unlike a regular car check-up that allows faults to be detected before a failure occurs", says Jacek Rafalak from Comarch, a company delivering advanced IT solutions, including mainframe environment support tools.

Health checks are particularly vital for the maintenance of large and complex environments that rely on mainframe computers with a DB2 relational database. "For a Db2, a health check involves an audit of the environment, which tends to be in constant stream because there is continual component migration and new apps are being installed; this means that the integrity and performance of the system may be compromised at some point ", explains Jacek Rafalak, Senior System Engineer at Comarch, recognize the title of IBM Champion for Analytics.

Mainframe environments are designed to resist a prolonged period of maximum load without their work being affected, but for this to be possible, they must be healthy. There should be no bottlenecks or weak points.

Health checks are usually carried out on mainframe environment especially for Db2 databases, but they may also cover DB2 subsystems, hardware configuration and backup environment settings for disaster recovery.

Health checks bring very tangible benefits, such as important savings and faster business processes. "Working with a branch of an international insurance company, we were able to extend its operating time by more than two hours per day, only because our analysis discovered that the disks needed to be replaced with faster models. This in turn made all processes that were dependent on data processing take much less time, which meant that the centers could be open longer", explains Marcin Trzaskowski, Data Center Manager at Comarch.

Health checks: how to maintain crucial parameters

The performance of any computer environment, including one based on mainframe and Db2 subsystems, will deteriorate over time; at some point, its configuration will no longer be optimal. This may slow down the system, cause problems in untypical situations and cause the environment to fall short of the required application parameters laid down in the SLA (service level agreement). A health check involves monitoring the condition of the system, as well as its metrics, recovery ability, DB2 performance and database engine. The audit, however, is not limited to problem detection, but also offers relevant solutions, based on the best practices established by Comarch.

Health checks look into the system's memory, processor (CPU) and database configuration, and track long transactions, locks and many other parameters that give a good picture of system health. Based on such analysis, recommendations are drawn up on what is non-optimal and should be changed in the system.

"The objective of a health check is to improve IT management, increase user satisfaction and identify and eliminate the system's weak points. However, it is a rather complex and expensive task that is burdening to the system, so it should never take longer than necessary", emphasizes Jacek Rafalak.

In complex systems, we must always make sure that a change in one element of the workload won't have a negative impact on the environment at large. "The larger and the more complex the system, the more important the health check", explains Jacek Rafalak.

When should a health check be performed?

You should consider a health check if, for instance, your data environment is very complex. "An audit of this kind will prove particularly helpful if the company is running low on IT resources. Equipment can be added indefinitely, but that's not the point. The key is to make the most of what you already have. When you reach the limits of your memory resources, need to increase the number of DB2 resources or install new, critical software, a health check will allow you to thoroughly check the functioning of the system", explains Jacek Rafalak. If the system is overloaded, you may reduce the workload, purchase extra hardware, or check whether the whole system has been configured optimally, and use the data to find new solutions within existing resources.

"It should also be carried out whenever you tweak database engine parameters or implement new software. You can perform it without a special reason as well, just to check the condition of the system. This may prove a source of useful knowledge from a business perspective", adds Jacek Rafalak.

You should also consider a system audit when planning to recoverability mechanisms; it will tell you whether the number of backup copies and disks is sufficient to support the operation. "Clients who have already created backup copies sometimes ask us to check whether their system would be able to recover within six hours. A health check is perfect on such occasions; it will verify whether actual recovery times are indeed consistent with the objectives defined in repair and recovery procedures", explains Marcin Trzaskowski.

How long does a health check take? How often should it be performed?

Since a health check involves a very broad range of tasks and applies to complex, advanced systems, it should not be performed too often or without a good reason. According to Comarch experts, the growing number of transactions and the increasing complexity of mainframe environments require that comprehensive audits should be carried out at least once a year (as long as none of the exceptional circumstances mentioned above occur).

Marcin Trzaskowski emphasizes that, as part of its capacity management service, Comarch provides 24/7 IT environment monitoring to detect alarming trends in monthly or quarterly cycles. "We carry out continuous analysis that may also suggest a need for a health check. For instance, performance issues may arise, or a client may report that data processing takes too long; an audit would then be required to determine the cause of the slowdown. A health check of this kind may help us ensure a faster app response", explains Marcin Trzaskowski.

A health check may be performed for regular Comarch clients or carried out as a one-off service. In the latter case, however, it would take more time. "If the capacity management service is already in place and the monitoring is on, we immediately know what's going on and the audit takes less time. But if monitoring is scarce and admins don't track critical resources, or do so inadequately, we need to turn on the monitoring and collect relevant data before we can perform a health check", explains Marcin Trzaskowski.

The time it takes to do a health check will largely depend on the complexity of the system. "If we know the environment and the number of interfaces is not huge, we can get it done in one or two weeks. But in large environments with complex processes, the estimated time would be closer to three or four weeks", says Marcin Trzaskowski.

It should also be kept in mind that, for mainframe solutions, often used in the finance and insurance sector and where processor availability is very costly, the service should be planned in advance to generate considerable benefits. "This is why I encourage all companies that rely on mainframe computers to work with us. Our analysts will perform a health check, verify the condition of your IT systems, and recommend the best solutions and optimizations to any problems that they detect", adds Marcin Trzaskowski.

How Can We Help? 💬

Want to reduce the cost of your IT infrastructure? Need improved data security? Let’s chat.

Schedule a discovery call