Skip to main content
Solved

What is your Object Storage Journey?


Hi Folks,

 

What has your object storage journey with Veeam been like and at what stage are you at now?

I started using S3 compatible the minute Veeam introduced it as capacity tier for Scaleout repos. I leveraged a 4 node Docker Swarm cluster with Minio. I chose Docker Swarm at the time as Minio advised to use containers for multitenancy and mentioned Kubernetes. I arrogantly thought I would read up over the weekend and on Monday get it going. That was not the case to say the least. Docker Swarm was much simpler and I eventually went with that. The solution worked well but my lack of deep understanding put some restraints on what I could do, not to mention that Docker Swarm, while being a solid solution did not have the incredible fine tuning capabilities that Kubernetes has.

Since then it has been all Kubernetes and S3 never far behind. It got to the point where I simply had to defect to the S3 compatible side of the force completely 😁.

 

So what was your journey, what are you using. What are the pros and cons of Object Storage in your views? 

 

 

9 comments

Userlevel 7
Badge +6

I am using Object First 😁👏🏻

Userlevel 7
Badge +19

As I’m a Customer, I don’t use Object Storage of any kind. I mean, I guess Customers do...it’s just I have no need to. I may see what vendors offer an OVA and test a few out.

I pose a return question Geoff 😊 ...what did you begin using Object Storage for? Was it just for labbing/testing? Was it an implementation decsion at your previous place of employment?

Userlevel 7
Badge +21

When Veeam first introduced this we had Hitachi HCP and I did some tests but that ended up being a disaster as it was not truly compatible.  Since then we have implemented things like Cloudian, ECS, and HCPCS which is compatible.  I also test Wasabi with our generous accounts we get to use through Drew and also will be testing OOTBI once I get it to our datacenter which is the one I am most interested in.

Scality Artesca also uses the SOSAPI as well which I put in the lab for giggles.  Our path as an MSP is moving more towards Object Storage as we need a way to better manage capacity where traditional block is becoming more a nightmare for us and object will be easier to scale for customers.

I am going to continue with Object moving forward and I am sure one day it will be the defacto standard.

Userlevel 7
Badge +6

We’ve started with Wasabi and are looking at Object First for our client base which requires a highly secure on-prem solution.

Userlevel 4
Badge +1

As I’m a Customer, I don’t use Object Storage of any kind. I mean, I guess Customers do...it’s just I have no need to. I may see what vendors offer an OVA and test a few out.

I pose a return question Geoff 😊 ...what did you begin using Object Storage for? Was it just for labbing/testing? Was it an implementation decsion at your previous place of employment?

I personally started to look into Object Storage a long ago, at the time when Ceph was not owned by RedHat. We even used it at the service provider I worked for because it was natively scale-out capable, plus all the cool things that makes it an ideal platform for storing large amount of data, like erasure coding, auto rebalancing of nodes when you add/remove them, multi-site deployments with replication. Admittedly, we used Ceph exposing still block volumes to Linux repositories, but when it then became a native target in Veeam it was a no brainer for many people to move to it, with still some huge caveats, main one being performance at scale especially when immutability is used.

Userlevel 7
Badge +4

Now I’m use Azure Blob Storage and Huawei S3 storage. I like that Veeam integrates with most S3 services.

 

Userlevel 7
Badge +8

Object Fist of course 😉 😁

I started with MinIO, the free appiance, then, Wasabi, now, OOTBI!

AWS S3 always was unreachable to me due to NDAs, budget, etc.

cheers.

Userlevel 7
Badge +8

Things operate very slow here. My journey is doing some testing to Wasabi at the moment and it is working great.  I did some testing on a few Pure arrays and had some really good results as well.

 

Userlevel 4

As I’m a Customer, I don’t use Object Storage of any kind. I mean, I guess Customers do...it’s just I have no need to. I may see what vendors offer an OVA and test a few out.

I pose a return question Geoff 😊 ...what did you begin using Object Storage for? Was it just for labbing/testing? Was it an implementation decsion at your previous place of employment?

Hi Shane, I (customer) am planning to deploy object storage not because the object storage concept itself but a low cost scale out CEPH based storage, to get very long retention periods for data in repo, with a low and predictable cost. When you’re running out of storage, you only have to ‘stack’ another server full of disks and expand the CEPH cluster on it. Like Amazon Glacier, but on prem. 

Comment