ProCare High Availability Management

ProCare HA Management performs daily checks of system health and HA well-being every 30 minutes allowing us to find issues as soon as they happen.

High Availability software brings many benefits to a business, the ability to have an up to date backup of critical data has a significant impact on the ability to survive an IT disaster.

Many benefits can be derived from managing planned outages and providing second copy data for test. With High Availability being undoubtedly a vital addition to any system within today's current environment, ensuring it is running smoothly is just as important. Time within a business environment is just as precious as the customers themselves and good HA management can take up a lot of it.

Many modern software products can automatically handle and repair data but there will always be occasions where human intervention is required to maintain a solid duplicate of data. Time spent ensuring your HA is running efficiently and effectively can prove very costly and very tedious through checking the same things day in, day out.

All software providers provide instructions for daily, weekly, monthly and quarterly checks to prove the integrity of the second copy. It is these checks which provide IT Management with confidence in their replicated data. As these are point-in-time checks, their validity starts to expire from the moment they are performed and it's for this reason that some organisations prefer to do these daily checks two or three times a day.

The frequency with which checks are made is a trade off between wanting to perform them repeatedly and the cost of man-time in doing so. At Chilli, the need for frequent checking at a lower cost was noticed. Was there a way we can check on a frequent basis without using significant man-time and consequently making the service too expensive for our clients? We simply wanted to provide a more proactive service at a lower cost.

The answer came in terms of automation, our consultants needed the information presented to them based on exceptions so they could focus on fixing issues rather than finding them. The software we have developed allows us to automatically start a 5250 session, navigate to the screens we wish to see and then pass that information through a rules engine which looks for issues and generates alerts.

This means we can perform daily checks of system health and HA well-being every 30 minutes allowing us to find issues when they happen rather than a few hours later. It also negates the human factor, it becomes tiresome to check the same thing day after day when it always reports as ok and experience suggests checks get dropped if errors aren't found regularly, software on the other hand is relentless.

The checks we perform incorporate all of the things recommended by the software vendors plus additional checks based on our experience of supporting the platform and it's products.

Chilli's HA maintenance service provides:

• Automated Monitoring and Alerting of System and HA conditions

• Experienced consultants available to provide advice and hands on fixes

• Significant savings on customer man-time effort

• Monthly statistical and configuration reporting

The monthly statistical and configuration reports have proved to be highly beneficial to our clients and we believe they add significant value to the service we provide. Below are a couple of example of the types of data our customers receive:

This example shows the exposed entries on remote journals within a business week (monday-friday). This is number of locally generated entries which have yet to be transmitted to the backup system. Ideally this should obviously be zero but due to the nature of communications there will usually be a low number of outstanding entries. High numbers of outstanding entries indicate that the bandwidth may struggle with peak workloads.

This example shows the maximum apply latency in a working week (monday-friday) on the HA system. This indicates the difference in hours between the current time and the time of the oldest entry that was generated. The graphs show the time delay experienced in the system which is split into time periods to highlight the time delay during specific job actions.