Subscribe

How to maximize Flash Pool capacity used?

Is there any way to allow Flash pool to use all of its capacity, at least for read caching?

 

This system is running ontap 8.3.1

 

FlashPool_FAS3250_CTL-7.PNG

 

Image from graphite+netapp-harvest

Re: How to maximize Flash Pool capacity used?

I have the same question about my environment.

 

Plus, I want to know what stats I can run to see this same information on the CLI that you are seeing in your graph.

Re: How to maximize Flash Pool capacity used?

Hi Nexica,

 

The maximum size of the Flash Pool cache per system is a function of the controller model and the Data ONTAP release in use. In addition, when both Flash Pool and Flash Cache are used on the same system, the maximum cache size for each caching solution applies, as does the total maximum cache value for the combined use. For details refer section 3.3 Maximum Cache Capacities of the TR-4070. You can refer Hardware Universe to get the specifications for the controller model that you are currently using.

 

For eg: I have pasted a copy of the limits for the FAS 8060 controller running cDOT 8.3.1. 

 

Flash.JPG

 

So even if your system has more cache, only the allowed maximum will be used. Hope this helps.

 

Regards,

Anoop

If this post resolved your issue, help others by selecting ACCEPT AS SOLUTION or adding a KUDO.

Re: How to maximize Flash Pool capacity used?

Hi, Annop,

 

The question isn't about the limits of flahpool. It's about how to maximize the usage of the currently installed (and within the limit) flashpool

 

Regards,

Re: How to maximize Flash Pool capacity used?

Hi,

 

Did you try to set the read and write caching for flashpool as per the caching policies defined in the section 2.4 of the same TR-4070?

 

Note: NetApp recommends using the default Flash Pool read cache policy. The Flash Pool cache dynamically allocates caching capacity based on the I/O demands put upon the volumes in the aggregate. With the exception of excluding one or more volumes from using the cache, the default "random-read" cache policy  should suffice for the vast majority of workloads.

 

Regards,

Anoop

If this post resolved your issue, help others by selecting ACCEPT AS SOLUTION or adding a KUDO.