Qlik Replicate Source db2 iseries cpu consume high

[lguevara]

Hi,

What configuration we do to reduce the consume of cpu in the source db2 i series,

because we notice that have to 10% of pick consume, but only have 1 task like log stream with 3 tables.

Thanks

[john_wang]

Hello @lguevara ,

Thanks for the post here.

In my view, a 10% CPU load on AS/400 isn’t directly comparable to the same number on Linux or other operating systems. For example, on Linux the CPU% reflects total cycles consumed across all cores. On AS/400, however, the CPU% may represent usage within a specific shared pool, and depending on how multiple threads are counted, it can even show values above 100%.

CPU usage on AS/400 can be influenced by many factors, such as the number of tables, the volume of change data, and the overall system configuration. We can certainly try to optimize things through Qlik Replicate settings or by using the R4I UDTF for Change Capture, but it’s difficult to predict exactly how much the CPU load will drop.

If you still feel that 10% is higher than expected, you can manage CPU consumption at the subsystem or job level. Just keep in mind that throttling CPU for the Replicate processes may introduce latency in your tasks — the classic “save CPU, lose speed” trade-off.

Hope this helps.

John.

[lguevara]

Hi @john_wang 

Thanks,

And How can manage CPU consumption at the subsystem or job level?

If increase frecuency of "Check for changes every (sec)" in advanced endpoint connection?

or another internal parameter.

thanks,

[AndyWaldrop]

Did you find an internal parameter to fix this, possibly around pooling connections?  We are having the same issue and Qlik is consuming up to 99% of the cpu.

[DesmondWOO]

Hi @lguevara , @AndyWaldrop ,

From my understanding, the CPU% is dynamically managed by the system. The IBM i (AS/400) architecture is designed to maximize hardware utilization; if no other high-priority tasks are competing for resources, a batch job will naturally scale up to use the available capacity.

Importantly, whether the CPU% shows as high or low, the actual CPU Time (the total work performed) remains nearly identical for the task.

If high CPU impacts Interactive Response Times for other users, an administrator can adjust the Job Priority (Run Priority) to ensure other tasks take precedence. Additionally, if multiple tasks are redundant— reading the same journals—combining them into a single task can reduce the overall processing overhead and lower the cumulative CPU demand.

Regards,
Desmond