Hello all
This case is about memory resource usage of the controller is overload in OceanStor 5500 V3.
【Problem Description】
Memory resource usage of controller 0A has exceeded the overload threshold.
【Problem Analysis】
In the storage log, we found both 0A and 0B controller has reported this alarm.
From the performance data, we found as below:
The Veeam luns took too much cache page resource. So if other services take a little more cache page, the memory threshold will be reached and the alarm will be reported.
Why the cache page perforation is so high for those veeam luns?
From the disk utilization statistics, all disks have high utilization. So we can judge that the service load is too large, so that exceed the disk bottle neck.
Storage handling IO process: host -> storage cache -> disk
If the disk utilization exceeds 70% frequently, then disk should be often
in a busy state. Then the IO will keep in the cache when the disk is busy. So the cache
will be taken a lot. 
We can see all veeam luns have a very large delay. But Veeam luns should take the backup service, may not have many requirements on the IO delay. So you did not find the service impact.
【Root Cause】
The service load is too large, so that exceed the disk bottleneck.
【Solution Description】
If there is no other important service or Services which is sensitive to IO latency. Then you can ignore. If the customer wants to resolve the performance issue, then need to migrate some backup service to another storage. Now this storage was almost in overload state all the time.
Thanks.
