I recently activated the Cockpit Server and Agent (4 SP08 PL1) to monitor a single SAP IQ server 16.0 SP8 LP3. I also set the alert to send me email using the default alert threshold.
After a few hours I started receiving alert from cockpit indicating that the main cache in use exceeded the thresholds. As the main cache was getting filled up, the alerts were changing from LOW to HIGH priority. At the end, I was flooded with alerts for the main cache and disabled the monitor for the main cache.
As per SAP IQ documentation:
“Number of buffers in use. Normally this is 100% because the buffer manager leaves buffers in memory until they are needed for some other data. In general, the buffers used and buffers locked numbers are meaningless, because SAP IQ uses buffers as aggressively and efficiently as it can.
This value grows to the maximum number of buffers that fit in the main buffer cache. The number increments whenever a buffer is allocated, but decrements only when a buffer is destroyed, not when it is unlocked or flushed. Objects in the temporary cache release their buffers when they are finished, but in the main buffer cache, SAP IQ may or may not destroy the buffers because as long as a buffer is unlocked, it is available for reuse, whether it is empty, contains data, or contains destroyed data.” SAP IQ SP08, Administration: Database; Table 27: IQ Main Store Blocks Message; Page 320
It seem to me that having a Main Cache 100% should not raise an alert, but the alert should be raise if the number of disk I/O is high indicating that there is excessive trashing for the main cache.
- Do I’m missing something?
- What is the use of having and alert to monitor the main cache used?
- Is there a method to flush the main cache while the database is online?
Thanks a lot,
Uvernes.