It’s on by default and there’s no straightforward way to disable it, so it’s likely on.
You can verify what it’s set to with:
::> set diag; node run * flexgroup set
That gives a list of current per-node settings. proactive-resizing is what you should see and it should be “true.”
I took a look at the ASUP from scarab. There’s definitely something wonky going on with that FG. No elastic sizing and no proactive sizing happening (note the green line is flat). And member 2 is way more full than the others.
So there may be a larger file in that one.
Have a look at this blog to see how to identify files in member volumes:
I’d suggest also opening up a support case.
From: Toasters
toasters-bounces@teaparty.net On Behalf Of Rue, Randy
Sent: Wednesday, May 26, 2021 3:03 PM
To: Toasters <toasters@teaparty.net>
Subject: Re: WAFL error on a flexgroup?
|
NetApp Security WARNING: This is an external email. Do not click links or open attachments unless
you recognize the sender and know the content is safe. |
Thank you.
I'm not finding any mention of turning this feature on or off, is the default? If so, we're at 9.8P1 and should be running it.
On 5/26/2021 11:44 AM, Parisi, Justin wrote:
TR-4571
https://www.netapp.com/us/media/tr-4571.pdf
From: Rue, Randy <randyrue@gmail.com>
Sent: Wednesday, May 26, 2021 2:43 PM
To: Parisi, Justin <Justin.Parisi@netapp.com>; Douglas Siggins <siggins@gmail.com>
Cc: Toasters <toasters@teaparty.net>
Subject: Re: WAFL error on a flexgroup?
NetApp Security WARNING: This is an external email. Do not click links or open attachments unless you recognize the sender and know the content is safe.
OK, after reading the below it sounds like elastic sizing should already be on and should have prevented my full constituent volume. Why didn't it?
What's proactive resizing? Where can I find out more about it?
On 5/26/2021 9:23 AM, Parisi, Justin wrote:
I’d actually recommend 9.8 and proactive resizing instead.
Elastic sizing is reactive and incurs a latency penalty when used. And it can’t be used with vol autosize.
Proactive resizing starts resizing when we get to 60% used and maintains a relatively even free space buffer and can be used with autosize.
From: Toasters <toasters-bounces@teaparty.net> On Behalf Of Douglas Siggins
Sent: Wednesday, May 26, 2021 12:15 PM
To: Rue, Randy <randyrue@gmail.com>
Cc: Toasters <toasters@teaparty.net>
Subject: Re: WAFL error on a flexgroup?
NetApp Security WARNING: This is an external email. Do not click links or open attachments unless you recognize the sender and know the content is safe.
I believe this might help:
Elastic sizing was introduced in Ontap 9.6. This article details the issue:
This should be turned on by default but can be turned off via:
node run nodename flexgroup set elastic-sizing=false
On Wed, May 26, 2021 at 12:01 PM Rue, Randy <randyrue@gmail.com> wrote:
OK, resizing the flexgroup added some space to the constituent volume
(to all of them, actually) and now it has free space. But a look at the
eight CVs in the FG shows they range from 22% to 68% full. Is something
keeping them more or less balanced? If so, how did it fail? And if not,
what the hell?
On 5/26/2021 8:47 AM, Rue, Randy wrote:
> sure enough, that constituent volume is completely full:
>
> scarab::> vol show -vserver scharpdata -is-constituent true
> Vserver Volume Aggregate State Type Size
> Available Used%
> --------- ------------ ------------ ---------- ---- ----------
> ---------- -----
> scharpdata scharp_systems__0002 scarab_b_SSD_1 online RW 448GB
> 0B 100%
>
>
> now what?
>
>
> On 5/26/2021 8:38 AM, Vanderyt, Jeff wrote:
>> vol show -vserver vserver_name -is-constituent true
_______________________________________________
Toasters mailing list
Toasters@teaparty.net
https://www.teaparty.net/mailman/listinfo/toasters