Volume Autogrow Changes in OnTAP 9.2

[TMADOCTHOMAS]

First, some context/background. We have a single 4-node cluster, all AFF, running OnTAP 9.2P2, with every volume configured to thin provision and autogrow (but not autoshrink).

I discovered today that the autosize-grow-threshold-percent volume setting is not fully conforming anymore to the standards discussed in these forums several years ago.

Specifically, the following post lists the point at which volumes of different sizes will trigger autogrow events. I have used this for years to determine when to alert on low volume space.

http://community.netapp.com/t5/OnCommand-Storage-Management-Software-Discussions/Controlling-autosize-threshold/td-p/70310  

The following KB article further clarifies that in Data OnTAP 8.2, those settings are still the same but administrators can now change them on a volume by volume basis:

https://kb.netapp.com/app/answers/answer_view/a_id/1028424

Today I ran the following command. Based on the above articles, I should have had 0 results (I have never manually chnaged the autogrow settings for a volume), however I found 74 volumes that were using autosize-grow-threshold-percent settings higher than 85%. Some were 90%, some 92%, some 95%. There were 127 volumes at 85% as I expected.

-----------------------------------------------------------------------------------

volume show -fields size,autosize-grow-threshold-percent -size <=20GB -autosize-grow-threshold-percent >85

-----------------------------------------------------------------------------------

Does anyone know if this behavior has changed in 9.2? I can't find a different version of the KB shown above for 9.2, or any other documentation about changes.

[davehahn]

I opened a support case about this issue and the support representative gave me the following table for default autogrow percentages based on the size of the volume:

[TMADOCTHOMAS]

Right, that's the exact list I've been using for years that's covered in the link in my initial post. The issue is that these percentages (in some cases) are different than they should be. Since posting this, I've discovered that resizing a volume doesn't also change the threshold. It stays the same based on the original volume size. Not intuitive at all but that's how it works. The only fix I know of is to audit volumes every so often and correct any that have changed.