convert flexvol to flexgroup ?

[regal swan]

Hello,
We have big flexvol volume that will probably reach 100TB in a couple of monthes. Knowing that 100TB is the maximum size of a volume, wich alternative do we have to avoid this situation ? I heard that Flexgroup could be a solution.
Is it possible to convert, or migrate a flexvol to a flexgroup ?
Any advice, best practice that could help us in our situation ?
Thanks a lot !

[somber birch]

https://docs.netapp.com/us-en/ontap/flexgroup/convert-flexvol-concept.html

it's doable. what ONTAP version are you on though?

best practice to move from one to another, it could matter.

[regal swan]

thanks !. We are in ontap version 9.9.1P6 . FAS8300 model

Is flexgroup working the same as flexvolume regarding the tiering of cold data to fabricpool ? We have storage grid system.

[somber birch]

yep, fully supported - re: during the conversion - check starting page 179 - https://www.netapp.com/pdf.html?item=/media/12385-tr4571pdf.pdf

FabricPool does not need anything. It just works. No need to rehydrate data on premises.

[worldly estuary]

For an offline conversion, you can do it on any current release - ie create a new flexgroup, take the source offline to users and copy the data using XCP or other. If you need to keep the source volume online, you really want to be on 9.12.1 so that you can auto-balance the volumes. If you are pre-9.,12, you can convert online but you'll have 1 constituent at close to 100TB and the rest of the constituents at 0 to start with.

[regal swan]

Hello,

I'm back with my flexvol to flexgroup conversion.

We have converted a flexvol to flexgroup and have the following situation :

• 2 constituents : one containing datas, the second empty

cluster::> volume show -vserver vserverA -volume-style-extended flexgroup
Vserver Volume Aggregate State Type Size Available Used%

---

vserverA volume_B - online RW 38TB 21.47TB 40%

cluster::> volume show -vserver vserverA -volume-style-extended flexgroup-constituent
Vserver Volume Aggregate State Type Size Available Used%

---

vserverA volume_B__0001 first_aggr online RW 24.35TB 8.50TB 63%
vserverA volume_B__0002 second_aggr online RW 13.65TB 12.97TB 0%
2 entries were displayed.

cluster::>

We have Ontap version 9.13.1P10

We have tried to rebalance the flexgroup but nothing changes.
Is the rebalancing feature corrected in a newer release ?

a couple of questions :

• what is the impact of keeping this configuration ?

• can we keep this flexgroup configuration like this ?

• Is this configuration common to all flexvol to flexgroup conversion ?

• when new data will be inserted in the volume, where will they go ? in the first constituent or in the second ?

• when will the second constituent start to receive datas ?

Thanks a lot for any advice,
Vhyse

[toxic bramble]

Rebalance is not intended for use after conversion. Best practice is to create a new flexgroup and copy the data from the volume in it.

This also gives you the option to use smaller constituents

Impact is that you only have extra space, no distributed performance.
You can keep it, bit it is not recommended.
Yes, this is very common, why conversion should be used with care, depending on your environment

If you just want space, this is fine. New data will be distributed. A balanced setup can take a long time. Depending on hot/cold data.

[sly hill]

@regal swan in later versions of ontap you can grow past 100tb volumes without the need for flexgroup. But you then need to think how long that would take to recover in dr scenarios. Same for flexgroups essentially anyway as they have limited dr options

[spare seal]

If only someone had blogged about this.... https://community.netapp.com/t5/Tech-ONTAP-Blogs/FlexVol-to-FlexGroup-Migrations-Things-You-Should-Know/ba-p/456417