Disk space per node

Hi, I have a question regarding space per node.
From the storj document it said the minimum req. space for node is 500GB which maximum is 24TB. If I currently operate the node at 24TB and planning to expand what should I do between expanding the currently node exceed 24TB or start an alternative node?

Ps. Base on same public IP environment.

I think 24 TB was supposed to be maximum recommended space, not maximum space but in any case there’s no indication of benefit of such large nodes for now.

Demand data is not yet available as we are only about to enter production now. You should have a better idea over the coming months.

If you are asking whether to expand the current node or start another node on the same address, the more nodes, the less the risk in case of some kind of hardware failure, so I would suggest multiple nodes if you ever run into capacity bottlenecks. In case you’re running some kind of fault tolerant storage system, I would say it’s about the same, I suspect Internet connection is the weakest link there.

The answer is simple: 1 node per Harddrive. If you use raid, you might use 1 node.
However there is no point in expanding a node if your used space is still below 70% and I doubt that you even have 10TB full yet.

Thanks for your advice!!

So multiple node (in the future) should be great idea than single node.

Thanks for your answer!!

My current node is on raid 5, and you’re right I still have a lot of free space for now.

Keep in mind that the multiple node advice is based on using different hardware for the other node. If you’re expanding the existing array and then run 2 nodes on the same array you see no advantage, only added overhead. I general I agree that 1 node per HDD is the best solution as you get to share all your HDD space and you have independent fault domains for at least the physical HDD. But since you mention 24TB and RAID setups already you might be one of the exceptions.

My particular setup is a NAS that is used for low impact usage. Some backups, streaming and iSCSI. The iSCSI’s are used with a large local SSD cache on the machines using them. In short, non of these uses require much from the resources of the NAS or speed of the array. So I use an existing RAID 6 array (technically SHR-2). In my case it makes more sense to run one node on the spare space on that RAID 6 and benefit from the dual disk redundancy without having to use more disks. The node also sees a slide performance increase from the 1TB SSD cache in the NAS. I you think will know best whether this applies to you as well.

1 Like

In my opinion, the uptime requirement and no way to back up the node means I have to use RAID.

1 Like

There are very good existing discussions on this topic of RAID vs 1 node per HDD. Since the opinions vary on whether the Storj recommended approach of one node per disk is the best one, I’ll just link two of them.