3PAR8200 Pre-update checks warning

[CRD]

Good day

Want to do some patching on the 3PAR8200 array but keep on getting the below when running the pre-update checks. Even during low IO time frames, is it safe to proceed with update? 

Any advice will be appreciated

Thanks

Node0 CPU load average is 65.69 percent.
Node1 CPU load average is 52.63 percent.
Due to high utilization on the array, the update may take more time to install. OS updates should be installed only when the average CPU utilization is less than 50 percent as measured by the CLI statcpu command

[Sheldon Smith · Accepted solution]

When you ran the pre-checks, you had roughly 60% load on each controller's CPUs. If you had gone ahead with the update, there would have been a time when one controller has to reboot with updated code. The remaining controller's CPUs would be at 118%. This would happen again later when the second controller has to reboot.

The system gives priority to data over background tasks (like updating controller firmware). Even so, with a peak(?) CPU load of 120%, when one controller is down, regular data IO will have some performance degradation. I don't think there will be a problem with the update completing, but as it noted, it will take longer than normal.

Look at the array's historic StatCPU performance data and see when there is a consistent time of day (of week?) when the total load drops below 100%. Preferably below around 85%. And do the update then.
Alternatively, you may need to plan on an outage of at least some of the workload to artificially drop the total CPU load.

The array has two resources: Disk space and Controller CPU. Just because you have free disk space on the array does not mean the array can handle the workload.
You need to consider moving some of the workload off the array to somewhere else.

[CRD]

Thanks Sheldon

Much appreciated and noted

Regards

[Sunitha_Mod]

Hello @CRD,

Excellent! 

We are glad to know the problem has been resolved.