it’s just expected, it’s not practice.
When you keep 10 nodes on the host (and there are 20 screws there), expectations begin to diverge very much from practice. you begin to understand this and begin to better understand the technical part of the project. When you have ten hosts with different processors and a disk system, your practice and understanding increases. When you read everything from hundreds of people, including their questions to the Labs and their answers, you begin to understand a little more. But not from everything, but only from what I remember, what was interesting to you.
People with really large volumes just don’t write anything here. And I probably won’t write like that anymore. Frankly and in many ways. This time is wasted.
Well, you try to make GE from one host with 5 and 10 nodes, then from another, well, let’s say the first was the old intel xeon, and the second was the new Threadripper. And compare.
You are a theorist. It’s not interesting to me at all. You will not tell me anything new that I had not heard before and had not discussed this ten times.
Your NAS is not accessable from the Internet.
“The connection has timed out”
Same here. The link isn’t working for me.
It worked fine for me a few hours ago. I’ve uploaded it to my personal server for those having trouble downloading it from the original link.
Yes, my internal DSM Firewall with country filtering rules was on. Glad to here that it worked to block all no-germans 
Now it’s open and accessable, but thanks to cdhowie for posting…
It didn’t work then; I’m in the US. 
I use RAID6 (well, raidz2) with 6 drives for the reason that I do not want to start over and wait a year until that node is fully back up.
Also, managing many nodes means I have more opportunities to make mistakes, even though the impact of a mistake would be lower.
There is no question that managing one large logical volume is far easier than managing a bunch of independant disks. I also am running raid6 on my older nodes.
Especially when it’s cheap 3tb drives (20$), and 10tb node.
$20!? Where are you getting those?
Probably used drives. Still, I am using RAID6 with new drives too, after a while, new drives become “used”.
Maybe I would run separate nodes if I could have backups of them or the uptime requirement was way lower. No backups + strict uptime requirement = too much risk.
I already answered you a similar question
backup or a raid solution is required to ensure SNO doesn’t loose withheld amount. If there was no withheld amount then you are correct no raid is required or backup solution since data is already redundant in the network.
Backing up a storagenode is an impossible task. Seconds or minutes after you made a backup it will be useless as there is new data in your storagenode. If you restore your backup you will fail audits pretty quickly. You’re better of creating a storagenode on that backup hard disk and earn money with it.
Backup is impossible. It’s RAID or no RAID, having an old copy of the storage node data is completely worthless.
despite backup being impossible: backup per definition means +100% disk space to be bought.
As I clearly show in the excel-calculation, even raid(5) does not make sense financially.
Have a look to the spreadsheet:
https://static.chrishowie.com/Raid-Comparison_v01.xlsx
But I also have to agree to joesmoe’s argument that one giant node that you take care of is way more easy for the average user to manage, update, control other than many nodes that always must be setup new, because of the disk failure. If Storj is your main business you probably want to have maximum financel benefit and dig into it 24/7.
Time is money. So. But less money is more time…
The average user (like me for instance) might be running the node on a more-bay NAS where also privat data and services run. Spending one bay per node and still have the privat data set on 2bays in a raid1-setup doesn’t leave too many bays for the nodes.
E.g. 4bay NAS, 4TB private data. Wasting 2bays on 2x4TB Raid1 leaving 2 bays for max. 2x16TB=32TB for nodes. Where a raid5 with 4x16TB ammounts to 48TB-4TB=44TB for the nodes.
An upgrade from a 4bay NAS to a 8bay NAS does apply to approx. +500€ of costs.
Every being must find the preferable way of share Storj-Space. More cost dominated or more work-life-balanced.
Your post has made me think, what if we could use dropbox for backup if we will ever be able to use backups?
Technically, if you get a business team license you have unlimited storage, at least according to this thread:
what if 3 people went together and got such a license? Each one of them could upload however much they wanted and that for a flat $15 per user per month.
Could be profitable in high capacity scenarios but quite the hack
Or spinning this further, what if we could make a node that stores the files on dropbox somehow and deletes them locally, i’m sure there would be problems regarding the immediate file availability and many more but i’m just pondering
There are two ways of doing backups. One way would be creating a snapshot - it protects from some problems, but does not take a lot of space. Another way would be to use network storage server (where I cannot run a node) or a tape.