Boot Menu Option 4

[TMADOCTHOMAS]

I am developing a plan to sanitize all disks on a 2220.  If I turn off failover and reboot each node, choosing option 4 from the boot menu, will it blow everything away and create a new 3 disk aggregate, or will it blow everything away and create a new root volume in the existing aggregate?  I really need it to do the former, but am not sure option 4 will work.  My goal would be to then sanitize the disks that don't make up the root aggregate, then create a new root aggregate/volume on the newly sanitized disks, and finally blow away and sanitize the disks that were used for the original root aggregate.

I was hoping to use option 4a, which spells out the scenario I want (blow everything away and create a new 3 disk aggregate) but this option has been removed in 8.1.2, which is the O/S in use on this 2220.

Does anyone have a recommendation?

[aborzenkov]

This option will zero out all disks assigned to a filer and create root aggregate with root volume in it.

[TMADOCTHOMAS]

Thank you aborzenkov! Do you know if it is a three disk root aggregate?

[TMADOCTHOMAS]

Also: do you know if the reconfiguration affects the service processor configuration? It would be great to be able to use the SP to complete the reboot and sanitize procedure, but I don't know if it gets wiped out as well.

[david_sole]

Hi TMADOCTHOMAS,

Option 4a from maintenance mode boot is present on a Data Ontap versions previous than 8 (for example 7.3.7).  In that versions you can choose option 4 (no root volume) or 4a (make a root volume).

In Data Ontap 8.x option 4 makes a new file system on a new aggregate (aggr0) that contains the root volume (vol0). This option needs at least 3 disks because RAID-DP (2 parity + 1 data).

As you asked this option not affects SP/BMC/RLM configuration. If you have an ip configured on SP module you can execute option 4 remotely.

David Solé Pérez

http://blog.davidsole.es

[TMADOCTHOMAS]

Thanks David!  So on our 2220, I have 9 of the 12 disks assigned to filer A, 8 of them making up the lone aggregate (aggr0) on RAID-DP and 1 for spare.  The B side has a RAID4 with 2 disks and 1 for spare.  If I choose option 4 on the A side, it should blow away my existing aggr0 and create a new three disk aggr0, right?  If so, my plan would be to sanitize the remaining disks assigned to A, create a new root aggr on the newly sanitized A disks, then blow away and sanitize the original 3 disk aggr0.  I would then assign any three of the sanitized disks on A to the B side so I would have enough disks to do the same thing on B.  Does this sound reasonable?

[david_sole]

Thanks TMADOCTHOMAS,

If I understood well you have a total of 12 disks in your storage system. The best choice in NetApp aggregates is to have all disks that you can assigned on the controller. In HA environments I usually configure the same number of disks on both controllers (in this case, 6+6) for best a performance, scalability and so on.

You can serve data from two controllers resulting in the use of all hardware (CPU, RAM, cache …) from every pair.

Initially when you start installing this type of controllers you lose about 50% of total space cause raid-dp and spare (I’m refereeing to 12 disks cabinets) but newly added disks are “pure data” (remember Mebibyte size, wafl reserve, snap reserve of aggregate and/or volumes).

If you want to initialize your storage controller (zero data) you can go throught option 5 and reassign all disks (6 + 6 for example). In my blog you can see how you can do it http://blog.davidsole.es/cero-del-sistema-de-almacenamiento/ (it’s written in Spanish but you can translate it) 

David Solé Pérez

http://blog.davidsole.es

[TMADOCTHOMAS]

Thanks David!

[aborzenkov]

Not in my experience. I performed reinstall via SP and at least until first reboot SP configuration was not affected.

[TMADOCTHOMAS]

Excellent, thanks aborzenkov!

[aborzenkov]

Yes, this is three disks aggregate (unless you have only two disks, in which case RAID4 will be created)