Object Storage Setup with Ootbi


Userlevel 7
Badge +22

Today we are going to setup Object Storage for our lab leveraging Object First’s Ootbi. This is a test lab environment for demo purposes. 

After powering on Ootbi we are brought to the license screen.

License

After accepting the agreement we land on the Available Options page.

Available Options

We will now setup a new cluster.

New Cluster Setup
Hostname
Cluster name and Cluster IP

Next we need to give our system a password.

New Password

We are almost done.

Completed Successfully

Turn on the Telemetry service.

Telemetry Service

We have now completed the Ootbi setup.

Cluster Setup

Next we can move on to the Web UI.

Web UI

After logging in for the first time Ootbi will give us hints of what our next steps should be.

Hint

We will now create S3 Access keys.

Create New Key
Keys Created

Now Ootbi wants us to create a bucket. 

Create a Bucket
Bucket Creation
Bucket Created

We are now ready to move on to Veeam. Remember to copy the endpoint for use when connecting from Veeam.

Endpoint

We will now add Ootbi as a Veeam Repository.

Add Backup Repository
Object Storage
S3 Compatible
S3 Compatible
Add the keys

You don’t need to write the keys down as you can always retrieve them in Ootbi.

Highlight the key and press show details to see both keys.

Keys
Enter the Keys
Find your Bucket
Create a Folder
Make recent backups Immutable
vPower
Press Apply
Repository Added
Finish

There we have it all done.

Repository Added

In a future post we will create some backups jobs and leverage Ootbi as the target taking full advantage of SOSAPI!


13 comments

Userlevel 7
Badge +20

This is awesome and I will be following this once I receive OOTBI 😁

Userlevel 4
Badge +1

Nice!

 

“S/N: VMware-56”, so it’s a Virtual Appliance? I understood they didn’t have one...

Do I need to bug a bit more my ex-colleagues?😝

Userlevel 7
Badge +22

Nice!

 

“S/N: VMware-56”, so it’s a Virtual Appliance? I understood they didn’t have one...

Do I need to bug a bit more my ex-colleagues?😝

Hi Luca,

I think you don’t even need to ask :). This is very new but there is definitely a plan to get one to folks like you.

Let me get back to you.

 

 

Userlevel 7
Badge +6

Nice!

 

“S/N: VMware-56”, so it’s a Virtual Appliance? I understood they didn’t have one...

Do I need to bug a bit more my ex-colleagues?😝

Hi Luca,

I think you don’t even need to ask :). This is very new but there is definitely a plan to get one to folks like you.

Let me get back to you.

 

That sounds amazing! 👏🏻

Userlevel 5
Badge +7

Looking forward to test it in our labs. this instruction will be helpfull

Userlevel 7
Badge +20

Would love to get an OVA format to test with too.

Userlevel 7
Badge +17

Nice, simple setup process @Geoff Burke . Great post!

Userlevel 7
Badge +8

I can’t believe how easy that is!!! Very cool. Thanks Geoff. 

Userlevel 3

I’m wondering...any way to have a geo-cluster with that ? Two sites replicated as with ECS or RING ? Or is it in the pipe ?

If you bring us that...That would be marvelous and really easy to push everywhere :)

 

Userlevel 7
Badge +22

I’m wondering...any way to have a geo-cluster with that ? Two sites replicated as with ECS or RING ? Or is it in the pipe ?

If you bring us that...That would be marvelous and really easy to push everywhere :)

 

Hi vChris, I don’t see this right now. One of the main ideas behind Ootbie is to keep it as lean and mean as possible, not just for performance but also for minimizing the attack surface. By leveraging the 3 2 1 rule you will have redundant data and Veeam will be aware of where the backups reside. 

Userlevel 3

Having a single backup immediately duplicated because it’s a geo-cluster is even simpler than doing a backup copy. The attack surface isn’t greater this way as you may use the same interface or an additional IP for that (assuming you embark the geo LB). But I get the “Lean and mean” state. I will look forward to object first around here :) Thanks for your answer.

Userlevel 7
Badge +22

Hi vCrhis, That is a good point you bring up. My other concern when Veeam is not aware is because I once had a major issue with something similar, i.e. background replication taking place and Veeam not being aware of the replication. When it came time to restore we had big issues. Don’t want to name the vendors involved :) or the setup which was not great. Again this was a one of situation and I have not seen some of the other solutions out there which people tell me work well. 

Either way keep these ideas coming as I am relaying back. In fact I am going to ask internally here again about this and see if it has been thought about and will get back to you.

 

cheers 

Userlevel 3

My other concern when Veeam is not aware is because I once had a major issue with something similar, i.e. background replication taking place and Veeam not being aware of the replication.

That’s the greatness in a S3 geocluster, as it’s the same bucket accessible from either side Veeam doesn't know that it’s a different cluster. And it’s the same data on both sides. And as you can write to multiple buckets, you can have multiple buckets on multiple clusters replicated and the load balancer, that’s a key point here, will direct you to the different clusters depending on your rules and you will benefit of the performance of all nodes during backup and restore in addition to the three copies with a single backup (perf tier - 1 copy, capacity tiers on S3 geocluster - 2 copies). We achieved some decent design with that and huge perf as it scales “infinitely”. Have a great day.

Comment