High Cache Miss Ratio on a volume in Active IQ UM

[heightsnj]

We have experienced a performance issue on a workload mostly read. What we can see the problem might be with High Cache miss Ratio in Active IQ UM on a volume involved. As attached graph, we can see the ratio at 70% in avg when the job was running. In addition, there were also a couple of it's clones were running read as well around the same time. 

With this high ratio, could it be the cause of the slowness? What the solution could be?

Thanks in advance for your advice.

[paul_stejskal]

We need to take a step back and figure out what the problem is and where the latency is seen. Once the bottleneck is diagnosed then steps can be made to fix it. This KB has a series of steps and links to other KBs which guide you through that process: https://kb.netapp.com/Advice_and_Troubleshooting/Data_Storage_Software/ONTAP_OS/ONTAP_9_Performance_-_Resolution_Guide

Follow that resolution guide and report back.

[heightsnj]

Yes, I have already performed steps in KB, and don't see any obvious latency on disks, network, or NVRAM. The only thing I can see is the high Cache Miss Ratio, so, I am just wondering if it seems normal or not, and if it could be the cause. 

Thanks

[paul_stejskal]

Hmm. Can you explain more what the performance issue is? CIFS share? Database? Specific application? CIFS? NFS? SAN? What are latencies when good vs bad?