FAQ: Overview of RAID Fast Zero feature introduced in ONTAP 9.4

Last updated
Save as PDF
Share
1. Share
2. Tweet
3. Share

Views:: 1,480

Visibility:: Public

Votes:: 2

Category:: ontap-9

Specialty:: hw

Last Updated:

Applies to

ONTAP 9.4RC1 and later
RAID Fast Zero for disks

Answer

How is Fast Zeroing enabled?

Fast zeroing will be available for systems initialized with ONTAP 9.4 and later.

Does an ONTAP upgrade to 9.4 enable Fast Zeroing?

No, upgrading from previous releases to 9.4 does not enable Fast Zeroing.
The storage controller MUST be initialized with ONTAP 9.4 and later to benefit from the Fast Zeroing feature.

What if my system was shipped with 9.4 and later and I want to disable Fast Zeroing?

The system must be re-initialized to disable Fast Zeroing.
If you like to have ONTAP 9.4 and later with Fast Zeroing disabled, then enable the following bootarg from LOADER and re-initialize the storage controller.

LOADER> setenv raid.enable-fast-zeroing? false

LOADER> saveenv

Can I use the above bootarg to enable Fast Zeroing on an upgraded system?

No, setting the bootarg alone does not enable Fast Zeroing.
The system MUST be re-initialized with ONTAP 9.4 and later.

How do I know Fast Zeroing is enabled on my system?

For systems initialized with ONTAP 9.4 and later, run the following nodeshell command:

::*> run local raid_config Maximum disks possible 144 Maximum disk objects possible 720 Maximum raid groups possible 400 Maximum aggregates possible 200 Maximum stripes possible 2048 Average raid group size 14 No. of disks in system 15 No. of groups in system 3 No. of aggregates in system 3 Maximum write disks active 720 Maximum write groups active 51 Maximum reconstructs 2 Maximum reconstruction size 64 Maximum disk copy operations 4 RAID buffer reserve 58481 RAID adjust buffer reserve 0 Average raid disk write queue length 50 Average raid disk read queue length   50 Average raid disk write transfer rate 50 Average raid disk read transfer rate 100 Number of RAID I/O threads      1
Fast Zeroing is Enabled

How do I know my aggregates are made up of Fast Zeroed disks?

Run the following nodeshell command at the diag privilege level:

::*> run local aggr status -r Aggregate aggr0_5a (online, raid_dp, fast zeroed) (block checksums) Plex /aggr0_5a/plex0 (online, normal, active, pool0)     RAID group /aggr0_5a/plex0/rg0 (normal, block checksums)       RAID Disk Device          HA SHELF BAY CHAN Pool Type RPM Used (MB/blks)    Phys (MB/blks)       --------- ------          ------------- ---- ---- ---- ----- --------------    --------------       dparity   0b.00.1P2       0b    0   1   SA:B   0   SAS 10000 55176/113000448   55184/113016832 (fast zeroed)       parity    0b.00.3P2       0b    0   3   SA:B   0   SAS 10000 55176/113000448   55184/113016832 (fast zeroed)       data      0b.00.5P2       0b    0   5   SA:B   0   SAS 10000 55176/113000448   55184/113016832 (fast zeroed)       data      0b.00.7P2       0b    0   7   SA:B   0   SAS 10000 55176/113000448   55184/113016832 (fast zeroed)       data      0b.00.9P2       0b    0   9   SA:B   0   SAS 10000 55176/113000448   55184/113016832 (fast zeroed)

How do I know my spares are Fast Zeroed spares or Physical Zeroed spares?

Unless manually physical zeroed, all spares on ONTAP 9.4 and above will be Fast Zeroed spares.
To check, run the aggr status -r or aggr status -s from the nodeshell with diag privilege.
If a disk does not have the suffix (fast zeroed) in the last column, then it is physical-zeroed.

::*> run local aggr status -s Pool1 spare disks (empty) Pool0 spare disks RAID Disk       Device          HA SHELF BAY CHAN Pool Type RPM Used (MB/blks)    Phys (MB/blks) ---------       ------          ------------- ---- ---- ---- ----- --------------    -------------- Spare disks for block checksum spare           0b.00.21P2      0b    0   21 SA:B   0   SAS 10000 55176/113000448   55184/113016832 (fast zeroed) spare           0b.00.23P2      0b    0   23 SA:B   0   SAS 10000 55176/113000448   55184/113016832 (fast zeroed) spare           0b.00.1P1       0b    0   1   SA:B   0   SAS 10000 801787/1642060800 801795/1642077184 (not zeroed) spare           0b.00.3P1       0b    0   3   SA:B   0   SAS 10000 801787/1642060800 801795/1642077184 (not zeroed) spare           0b.00.15P1      0b    0   15 SA:B   0   SAS 10000 801787/1642060800 801795/1642077184 (not zeroed) spare           0b.00.17P1      0b    0   17 SA:B   0   SAS 10000 801787/1642060800 801795/1642077184 spare           0b.00.19P1      0b    0   19 SA:B   0   SAS 10000 801787/1642060800 801795/1642077184 spare           0b.00.21P1      0b    0   21 SA:B   0   SAS 10000 801787/1642060800 801795/1642077184 spare           0b.00.23P1      0b    0   23 SA:B   0   SAS 10000 801787/1642060800 801795/1642077184 ::*>

You can also use other nodeshell raid level commands to identify the FZ details on disks.

What if the spares shows 'not zeroed' with ONTAP prior to 9.4? When will the spares be zeroed?

That's not a issue.

There will be a delay as the disks need to have a background zero operation before they can be added to an existing aggregate or used to create a new aggregate.

How do I physical zero my Fast Zeroed spares?

A new diag variable -use-physical-zeroing is added to the command disk zerospares.

But in order to physically zero a Fast Zeroed spare, the spare must be in 'not zeroed' state.

If a spare is in 'fast zeroed' state and if you run disk zerospares -use-physical-zeroing true, then the command does nothing on the spares with ONTAP 9.4 and above, since ONTAP does not need to zero the fast zeroed spare.

Is upgrading from ONTAP 9.3 to 9.4 have any restrictions due to Fast Zeroing

No. Upgrade will continue but Fast Zeroing will not be available.

Is downgrading from ONTAP 9.4 supported?

Yes, downgrading from ONTAP 9.4 to 9.3 is supported whether the storage controller has Fast Zeroing enabled or not.

However, downgrading further from 9.3 to 9.2 is not supported with Fast Zeroing enabled aggregates. The following error message will be printed:

::*> system node revert-to -version 9.2 -node node1 -check-only true Error: command failed: The revert check phase failed. The following issues must be resolved before revert can be completed. The following aggregates contain raidgroups with fast zeroed disks: aggr0_5a, aggr0_5b, aggr1_5a, aggr2_5a. Each node containing a root aggregate listed above must be reinitialized before retrying the revert command. Delete any listed data aggregates using the "storage aggregate delete" command.

Are there any issues with partition-enabled disks and Fast Zeroing?

No. Fast Zeroing works with partitioned disks just like regular, whole disks.

Where is the Fast Zeroing information stored on the disk?

Each Fast Zeroed spare must maintain its own FZ signature in a persistent area.
So the information is stored on the disk label.

Can recovery tools like wafliron, aggr undestroy, rebuildtree, buildtree, label edit be used?

Yes. No changes with the recovery procedures.

Any changes with SnapLock aggregates and disks due to Fast Zeroing?

No changes with current implementation.

Is full disk encryption (NSE, SED, FIPS) supported with Fast Zeroing?

Yes, full disk encryption (NSE, SED, FIPS) is supported with Fast Zeroing.

Are there any changes in AutoSupport with Fast Zeroing?

No. But the "SYSCONFIG-R" and "AGGR-STATUS-R" sections in AutoSupport will contain Fast Zeroing information.

Applies to

Answer

Additional Information