New setup on existing raspberry pi 4 NAS with OMV

Hello everyone,

I am making a NAS with a raspberry pi 4 and a gigabit fiber connection for personal use. I wanted to get a 8TB drive for my personal files but I can get a good deal on a 12TB drive today. Is it possible to partition the drive and only use 4TB for storj while also running open media vault for the other 8TB? I also plan on buying a second identical drive and using RSync to backup my files, would I be able to partition 4TB of this second disk aswell and have a total of 8TB for storj ?

Thanks

Hey,

yeah…all your suggestions are possible without much hassle.
Just install Debian Buster, OMV and from this point on, it’s easy.
You install docker, portainer, watchtower as containers…mount your hdds/ssds as you wish, then install storjnode as a docker with the flag "-e STORAGE=“4TB” and your set.

If you need help, just reply.

Ok thanks! I didn’t know if you had to use whole drives for storj or if it was possible to store my personal data aswell!

Looks like I will get the 12TB drives then. They are an extra 100€ compared to 2 8TB drives but I figured I could get that back after a year with storj and make profits from then on?

There are different approaches also…you can partition the hdd to “virtually” split the hdd for storj and your personal stuff, OR you just have one big partition and use different folders. That’s a personal taste thing :wink:

Ok! That sounds good. Do you think that I will recup the 100€ in a year with 8TB shared ? The spreadsheet calculator seemed to say so.

Ohhh…I cannot answer this question, without many assumptions…
It depends on so many factors…like your ISP, how reliable your hardware is, how much data will be transferred via the Tardigrade network.

Example:
My first node (after some testing-nodes) is from “2020-06-16” and currently has like 5 TB stored. The last months it earned like $15 per months but it still only pays 50% of it. So…the bigger earnings will happen after the “held time frame”.

1 Like

Ok, thanks for the help!

Yes, you can. I recommend using LVM, as then you’ll be able to somewhat smoothly resize the partitions when necessary. Though, does OpenMediaVault require a separate partition? You don’t have to have a partition exclusive to Storj.

1 Like

No, OMV doesn’t require a seperate partition. I don’t think I will bother having multiple partitions then. What would be the best way of using 2 disks? I saw that most people set up 2 nodes but does that have any downsides? I have just ordered the 2 12TB drives! Thanks.

1 Like

For nodes, two (or more) separate nodes. For OMV you might want some redundancy though?

setting up two nodes has the advantage of putting less load on either one, due to the fact that ingress is split between them, and ofc that reads in theory would be split between them, ofc that won’t always be the case as reads must be considered random.

running two nodes with a drive each then would in theory make the drive which OMV runs on less busy and should then run better… tho streaming media might not be to heavy on a hdd, so maybe it doesn’t help anything.

the other option is a mirror, ofc this reduces your capacity and write by 1/2, but retains your read for both drives combined, this is a very performance and redundancy advantageous setup, but also one of the capacity wise least efficient in the raid world.

not really familiar with OMV so i cannot say if that would require redundancy or not, active data does tend to degrade and hdd’s are “fairly” unreliable media, if you plan to store important non recoverable video you should most likely use a 123 backup strategy and then ofc having the active data on a mirror doesn’t hurt.

on top of that you still cannot rely on the data being safe because of bitrot, in theory you could identity it and remove it by comparing all the stored data, but running something with built in checksums to easily and quickly identity problems is much preferable.

data storage quickly gets complex, just remember hdd’s or even ssd’s are not reliable storage and can essentially fail at any time, even when brand new…
ofc if the data isn’t important those kinds of issues becomes less of an issue.
but if you plan to sort and store stuff long term avoiding corruption is pretty nice, because it’s very difficult to remove after it’s gotten into your dataset.

even if you are just using it for keeping tvshows or movies, then if you spend any work on sorting it or making it work, that work will easily be wasted when it starts to go bad…

ofc old media tends of simply become so poor quality that often one ends up deleting to download better quality version years later… so it’s highly dependent on what you want to do, but there can be many considerations…

most likely you will get most bang for your buck, with a storagenode on each hdd and then OMV running from one.

I agree with @peppoonline: Answering this question is beyond tricky…

  • Frist, the estimator is… an estimator. Depending on real usage of the Tardigrade network, things could be better, things could be worse…
  • Second, payment are made in crypto assets, so the amount of money you will have after one year depends on if you convert your STORJ to fiat every month, or if you keep all your tokens and convert them after one year, in which case it will highly depend on the change in value of the STORJ token. If it went up, good news! If it went down, you’ll have lost some money along the way.
  • Then, depending on the tax system in your country, converting your STORJ to fiat may cost you some money again. In my country for instance, 30% (!) is taxed by the government, unless you’re under some threshold.
  • Also, depending on your hardware setup, you may want to take into account that your machine is consuming electricity which costs money to be substracted from your revenues (but as you’re setting a NAS anyway, maybe this can be ignored)

Finally, I find that running a node is not completely a “set it and forget it” thing.
The software is getting better after each release, but to make sure it runs fine and doesn’t get disqualified it still requires a bit of monitoring an time to investigate issues when they arise :slight_smile: