FAS2020 dual controller with real passive configuration

[#1]

Is there a way to configure FAS2020 with :

All disks assigned to primary controller to maximize storage

Secondary controller only there in case of falure of primary one ,

so active / passive configuration without ontap (and then disks) running on second controller ?

Even a solution without automatic failover and with service disruption can be a solution for me,

my primary goals are

     - to maximize storage

     - and when a failure occurs, to have a procedure to restart (even manually) the FAS2020 on the second controller

Thanks for help

[#2]

Hi and welcome to the Community!

Is there a way to configure FAS2020 with :
All disks assigned to primary controller to maximize storage

Nope, this is not doable - you need minimum 3 drives assigned to each of the controllers.

This post describes the 'active/passive' config:

http://communities.netapp.com/message/42579#42579

Regards,
Radek

[#3]

Hi,

thanks for your answer,

that’s what i read often in posts,

but I’ve heard about kind of solution :

Disabling CFO

when a problem occurs on first controller, starting manually the second one to do the job

anyone have try this solution ?

[#4 · Accepted solution]

This is possible, but honestly not worth the effort.  Worst case scenario is the system board fails and you're down for up to 4 hours until a replacement is sent out (unless you only have Next Business Day).  It could take longer than that just configuring and getting clients to connect to the inactive controller.

When the active system fails you will have to move all network connections to the inactive controller and reassign ownership of all the disks to it.  All of your clients will have to reconnect to the new controller because it will have no session information at all (the reason most customers go for at least the active/passive setup).  If you have FCP/iSCSI clients you have even more issues getting them to connect to a system likely to have different WWNN/WWPNs.  Once you have this working, you could set aside the replaced controller to handle another failure.  The problem there is your AutoSupport history is tied to the system serial number.  Every time you have a system failure, your history starts over again which could get confusing for NetApp Tech Support.

Unfortunately, there is no middle ground.  Either a system is completely standalone and you accept downtime for failures and maintenance or it's an active-active configuration that consumes drives, but significantly reduces downtime.