RE: Runaway FlexGroup constituent

That is some pretty wild imbalance.

Got a case open? Or at least a serial number for the cluster?

From: Scott Eno <[email protected]>
Sent: Thursday, December 17, 2020 9:05 PM
To: NGC-tnaple-berkcom.com <[email protected]>; Toasters <[email protected]>
Subject: Re: Runaway FlexGroup constituent

Hi Tim,

Thanks for the link!  It looks as if this bug wasn’t fixed until 9.7 so it very well could be what I’m hitting.
On December 18, 2020, Timothy Naple <[email protected]<mailto:[email protected]>> wrote:
Scott,

It is hard to say because this does not specifically mention 9.6P5 but I wonder if you might be running into this bug:

https://mysupport.netapp.com/site/bugs-online/product/ONTAP/BURT/1262593

I am using FlexGroups quite a bit (mostly nfs, but some cifs) and I have not seen this sort of imbalance.  I also wonder if something temporary is happening on the filesystem like some sort of large backup file being written.  In theory elastic sizing would cover for it as long as autogrow is off, but I am not sure I would want to get to 99% capacity on that one member to test that theory!

Thank you,
Tim

________________________________
From: Scott Eno <[email protected]<mailto:[email protected]>>
Sent: Thursday, December 17, 2020 5:21 PM
To: Toasters <[email protected]<mailto:[email protected]>>
Subject: Runaway FlexGroup constituent

Hi,

I have a troublesome volume in a FlexGroup that continues to fill faster than the other constituents.

 6.25TB 4.22TB  32%
 6.25TB 4.19TB  32%
 6.25TB 328.9GB  94%
 6.25TB 4.24TB  32%
 6.25TB 4.20TB  32%
 6.25TB 4.20TB  32%
 6.25TB 4.23TB  32%
 6.25TB 4.24TB  32%

With autogrow disabled, from what I understand after reading Mr. Parisi's document, elastic sizing will make sure writes will happen. This FlexGroup is a CIFS share for a document management system, so it's millions of little files.

Am I safe to leave the FlexGroup like this until the other constituents catch up, or is there a need to grow the entire FlexGroup?

Cluster is 9.6P5.

_______________________________________________
Toasters mailing list
[email protected]
https://www.teaparty.net/mailman/listinfo/toasters