SSD Sanitization Time?

A follow up to my January 2019 post - Disk Sanitization Time? - but this time we have SSDs!

I actually thought SSD sanitization time would be really really quick, but it wasn't.

I had to sanitize 2 x AFF A300 HA-pairs each with 72 disks (3 shelves) of 3.6TB SSDs.

The disk sanitization command I used in maintenance mode was:

disk sanitize start -p 0x55 -p 0xaa -r -c 3 DISKNAME

This uses:

  • 0x55 hex byte overwrite pattern for the first overwrite.

  • 0xaa hex byte overwrite pattern for the second overwrite

  • And a random hex byte overwrite pattern for the third overwrite

And the -c 3 means we have 3 cycles of this (so essentially blocks are overwritten 9 times.) 3 cycles are required to satisfy US Department of Defence disk erasure standards.

It took around 12 hours per HA-pair (I didn't hang around to watch it all complete), so you could say 4 hours per cycle.

I should point out that I had all my disks assigned to just one of the HA-pair controllers, and ran the disk sanitize on all disks at the same time (which you can do in maintenance mode, definitely in 9.7P16 which I loaded on the controllers just for the purpose of sanitizing them.) Potentially having sanitize running from both controllers at the same time would have been quicker, but one of the controllers on the second HA-pair was broken anyway, so I could not test this.

The KB which covers the process is here:

How to sanitize all disks in ONTAP 9.6+ - NetApp Knowledge Base

I did not quite follow the procedure above, and due to a broken controller on the second HA-pair, I had to be a little inventive. My procedure was this:

  • Halt both controllers,

  • Use boot menu and option 7 to load 9.7P16 on the controllers.

  • Boot into maintenance mode.

    • Aggregate offline (all aggregates)

    • Aggregate delete (all aggregates)

    • Make sure all partitions are assigned to one node.

    • Unpartition all disks

    • Make sure all disks are assigned to one node (if you have problems with disk reservations, you can just pop the disks out the front and reinsert it, that'll get rid of the reservation.)

    • Halt the node that has no disks assigned to it (otherwise that node will get spammed with errors in maintenance mode when you start the disk sanitizaion process.)

    • Run the disk sanitize command on all disks.

  • And wait!

Note: The thing with SSDs is that not all SSDs support disk sanitize. See:

PS Apologies, I had to disable comments because I was getting way too much spam. If you want to contact me, please see: Cosonok's IT Blog: About