David Hallberg

Remedies for RHOs—Part 6: The OS Path for Rapid Transactions

The first five installments in our series on Millennium® performance issues for RHOs walked you through the necessary steps for evaluating connectivity issues as well as queuing problems involving Websphere MQ, shared service proxies, interfaces and Oracle. This week I’ll finish up with yet another area where queuing or waits can be a problem: the operating system (OS).

The OS runs the hardware for Millennium and is often referred to as the backend or backend servers. OS capacity issues can cause Millennium to perform slowly or, worse, sporadically. A sporadic manner could mean that at 7 a.m. Millennium is working fine, but from 9-11:30 a.m. it’s slow.

Before detailing helpful OS diagnostics, I want to answer the larger question, “Why do I need to bother with all this diagnosis? I’ve got an outsourcer to troubleshoot for me.” Experience has taught me that any time you outsource or out-task work, you should audit it. That way, if your clinicians and business staff are complaining about performance, you will have the information you need to have a data-driven conversation with your outsourcer.

The generic complaint of “Millennium is too slow” will generally be met with deaf ears. A comment such as “I have a Run Queue on all application nodes exceeding 4x my physical CPUs” is something your outsourcer can track down and help you correct. So I urge you to maintain technical knowledge in-house or find a trusted consultant who can assist in gathering and interpreting key data concerning your Millennium operations.

So what can you do to determine if your application nodes have performance or capacity issues? Click here to read more.

Prognosis: Monitoring OS capacity and taking action to reduce or eliminate issues as soon as they start—instead of days, weeks or months later—will improve Millennium’s consistency.