Connecting a NetApp FabricPool to an external S3 target

By | 4. June 2024

As promised in the last article, here are the instructions on how you can connect an external S3 StorageGRID (or other S3) to a local ONTAP system for the storage of cold data.

The function behind this is called FabricPool by NetApp. The local RAID, in NetApp speak aggregate, is extended by a “cloud part”. Please read the previous sentence again. Yes, the RAID is extended by a cloud part!

As we old storage hands now all know…RAID (Redundant Array of Independent Disks) everything can go in nothing can go out, unless a disc fails and a rebuild has to be done.

So if you add a cloud tier to your local RAID (aggregate), this “cloud part” is added to your RAID. “Just” removing the cloud part is then no longer possible. So please think carefully before adding a cloud part to a RAID.

Have you already considered that it might be wise to move cold data to a cloud tier or have you sought advice? Are you aware of all the consequences?

Then let’s get started…

In your System Manager on your NetApp system. Click on Storage (1) -> Tiers (2) -> Add Cloud Tier (3)

And this is where the crux begins. What do I use now? NetApp is of course interested in you as a customer keeping your data on NetApp systems. That’s why setting up a FabricPool, i.e. adding a cloud tier to NetApp storage, is free of charge for you. So if you have another ONTAP system that is equipped with cheaper HDDs or QLC SSDs, you can store your cold data there free of charge using FabricPool technology. This is because ONTAP has also been “speaking” the S3 protocol for some time now.

Or if you have a NetApp StorageGRID (S3 object storage from Netapp) in your basement, you can also outsource to it. However, as already described in the last article, the StorageGRID S3 can also be located with an external provider. You can then also outsource your cold data there without having to operate the storage yourself.

However, if you want to move your data from your on-premises NetApp storage to AWS, GCP, Azure, IBM or Alibaba Cloud, you will need volume-based licences from NetApp to do this. $$$

But now let’s get started. Click on Cloud Tier, then on StorageGRID…

…then enter the access data for StorageGRID.

You can choose any name you like. The server name FQDN is given to you by your provider, you generate the access keys and secret keys yourself. You choose the bucket name yourself.

Further down in the menu, you will be asked which aggregate you would like to attach directly to this cloud tier. We won’t select anything here for now.

You will then see in your System Manager that a cloud aggregate is listed next to your local aggregates.

Nothing important or bad has happened yet. We have only added one cloud aggregate or cloud tier so far. You can see this here, as “mycloudtier” which we added last.

Now it’s time to get married

Now it’s getting serious…! First I click on the three dots (1), then on Attach cloud Tiers (2)

I say “Save”. This marries my local aggregate with my “mycloudtier”.

And then I get another warning as described above. Once in RAID always in RAID.

Of course I click OK.

And then I can go and set new cloudtier settings for each volume…

In this case, I only want all data older than 14 days to be swapped out.

Cool thing, Fire and Forget.

This means that all cold blocks that have not been used for more than 14 days are moved from my primary storage system to my cloud tier for cold data.

I hope this blog has given you some new ideas.

Best regards

André Unterberg aka DerSchmitz

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.