ONTAP Discussions

Ways to reduce Filer high CPU utilization

SRINIKALYAN
4,816 Views

Hi,

I am interested to know the number of ways to reduce the Filer CPU utilization. Most of the busy times, CPU utilization is above 90%. Customer is asking for the ways to reduce the CPU utilization below 60%. What are the performance metrics to check? We have access only to Operations manager and Netapp Management Console - Storage performance tools not the command line.

Please advise.

Model Number: FAS6070

Operating System: 7.3.4

Thanks,

Srini

1 ACCEPTED SOLUTION

lueth
4,816 Views

HI Srini,

I'd recommend not worrying too much about utilization unless the customer is having performance issues.  Besides the customer's workload, Data ONTAP has several background tasks that will be opportunistic and take CPU cycles.  The caveat is if this is an HA system then in the event of takeover the customer will experience a performance hit, and if this is the case it makes sense to try and get system utilization closer to 50%.  The easiest way to do this is by reducing the amount of workload on the controller but that's generally a non-starter.  Using the tools mentioned you could look at scheduled events like SnapMirror replication to see if you can move those to off peak hours.  The other suggestion is to look at the system in MyAutosupport to check the general health and see if it reports recommendations that could help improve performance.  And keep in mind you can always open a support case to see if they can help you here.

Take care,

-chris

View solution in original post

1 REPLY 1

lueth
4,817 Views

HI Srini,

I'd recommend not worrying too much about utilization unless the customer is having performance issues.  Besides the customer's workload, Data ONTAP has several background tasks that will be opportunistic and take CPU cycles.  The caveat is if this is an HA system then in the event of takeover the customer will experience a performance hit, and if this is the case it makes sense to try and get system utilization closer to 50%.  The easiest way to do this is by reducing the amount of workload on the controller but that's generally a non-starter.  Using the tools mentioned you could look at scheduled events like SnapMirror replication to see if you can move those to off peak hours.  The other suggestion is to look at the system in MyAutosupport to check the general health and see if it reports recommendations that could help improve performance.  And keep in mind you can always open a support case to see if they can help you here.

Take care,

-chris

Public