I have setup a node that is LIVE with 6.5 TB ( I could only select one drive in the setup ) but what I really want is to setup 4x6.5 TB and 2x10TB aswell.
How can I add more TB to my node?
Note: I tried to download Docker but i see the “System Requirements” for docker is
Windows 10 64-bit: Pro, Enterprise, or Education (Build 15063 or later).
and I only have a normal Windows 10.
I want to use VPN on my windows, but each time I connect to VPN my node goes offline. Do anyone know how I can run a NODE + VPN ?
Hi @ARA welcome to the forum. As of The V3 network, you can only host ONE hard drive per node, In the V2 network you could, but the feature was removed as it slowed the network significantly… I believe you can use a VPN when you are running your node as you need to be connected to YOUR network… I think.
You should be able to do the CLI install on Windows 10 Pro, but you can only host ONE drive per node with docker as well…
In order to have multiple nodes, you need multiple computers. for example, you need two windows 10 machines to run 2 nodes. For each node you need separate hardware…
It is possible to run multiple nodes on the same computer system using docker containers. However, if the computing platform is resource constrained, running multiple nodes will probably not lead to a good result.
Do we know why they want to do it like this, not allowing to add more TB? This mean that I can only use 1/10 of my TB for Storj NODE. ( I thought the network benefits of we adding more TB to it )
Are we 100% sure it is not possible to add more TB? I cant understand the logic there.
Hey everyone Brandon the Product Manager @storj here. Thanks for submitting this idea, Chris VdG has a point about the redundancy between raid and erasure coding. Our recommendation is to run one node per HD. This will actually spread your overall risk of a HD failing and losing the withholdings for that specific node ID. If you have 4 2 TB HDs then the best thing to do is run 1 node per HD and if one of those fails then at most you would only lose 25 percent of your overall withholdings amongst all 4 nodes. Instead of using raid to backup one of your nodes its may be in your best interest to just allocate that extra raid space to the network and profit off of it instead.
Yes, that is correct. Storj does not recommend it though, because it lowers the life expectancy of your hdd’s. (Depending on what RAID setup you configure)
To be fair, I don’t think a 6.5TB node will fill up soon. I’ve been running my node since March and it holds 3.5TB atm. I have 10TB shared. I think it’s best to just start with the single disk node. Perhaps by the time it does fill up there is a better way to share multiple disks on windows. Since you don’t have 10 pro, the docker setup isn’t an option for you. So yeah, see how it goes with your first node and worry about adding more when it becomes applicable.
You have to remember that internet speeds are more important than hosting 10 TB worth of storage…
@Enzo Upload speed goes over Download speed right? also what would you say is good internet speed?
@BrightSilence I understand, I thought they where filling up faster.
I think I will do that, wait til the HDD is full then add more TB. ( when that time comes I will use RAID on my disks )
also do the node need to be full to get the max reward?
There is a lot of test data on nodes that is now being removed as customer data comes in. So yes actual data is going up but storage used on nodes seems quite stable. At least in my experience.
I have made 3 nodes on my windows setup, no doker, but you need to know how to install Service manualy, i used SC.exe to install service, then you can make as many as you want, but behind 1 static ip your nodes wont get more data, it will be spreaded between your nodes, BUT it will give more realibility as if hdd crash only 1 node data will be lost.
Before i had 5hdd in striped it work very fast, but 1 hdd falure will kill all 13TB node, taday i devided node to parts.
My node has seen continuous bandwidth for long periods of time at 6 Mbps “Download” and 5 Mbps “Upload” simultaneously. As of the moment I’m posting this, I’m seeing 12 Mbps aggregate. I have also noted a few very short term spikes in bandwidth usage up to 81.08 Mbps aggregate. So, at 15 Mbps, the connection will probably be a little slow under expected production level loads.
You should not use RAID for storj nodes unless you are sharing free space on a RAID volume that already exists – that is to say, you shouldn’t create a RAID specifially for running a storj node. Let’s look at the various RAID levels:
RAID0 (striping, no redundancy): If any disk in the array fails, the whole array fails. The expected failure rate of the array is the sum of the failure rate of each individual disk. If a disk dies, you have to recreate the RAID0 and start your node over from scratch with a new identity and no stored data and the risk of this happening increases with each disk you add.
RAID1 (mirroring only): A waste of space; the network already has built-in redundancy. It would be better for you and the network to run a node per disk instead of using RAID1.
RAID5 (striping with 1x parity): This is the only RAID level I would consider using to run a Storj node. At the cost of one disk’s worth of capacity, you protect all the other disks in the array. Depending on how many disks you are using and their expected failure rate, it could be beneficial to avoid having to recreate nodes each time a disk fails.
RAID6 (striping with 2x parity): Probably a waste of the space needed for the extra parity stripe. RAID5 would be better.
RAID10 (striping over mirrored pairs): Wastes the same amount of space as pairs of RAID1 arrays but provides extra performance. I don’t recommend this for Storj for the same reasons I don’t recommend RAID1; run a node per disk instead.
Windows has the concept of “spanned volumes” which are basically the same thing as Linux LVM volumes that span physical volumes, or Linux md-raid in linear mode. They have the same failure characteristics as RAID0 and none of the performance advantages. The only advantage over RAID0 is that the physical volumes don’t have to be the same size.
tl;dr: Run a node per disk. Possibly use RAID5 if you understand the trade-off you’re making: you use one disk as an insurance policy against your time restarting nodes from scratch as disks fail.
I’m currently developing a network simulation based on payout of single drive nodes that fail and become disqualified versus nodes that run RAID with lower overall disk space but much higher reliability.
The problem with the above quoted statement is that actual storage space is only paid out in $1.50 per TB… while Bandwidth is paid out in $20.00 per TB… So, let’s take an example of a node that runs 4 TB in RAID 5 with three drives versus 3 nodes that run 1*2TB drives each. That’s 4TB total storage versus 6 TB total storage… or in terms of dollars:
At 100% capacity the difference in payout would be $3.00 per month. But if all the nodes are located behind the same IP address, the overall bandwidth is shared and does not increase will multiple nodes running.
So… what happens when the nodes fill up slowly rather than quickly?
Ans: The payout will be the same until drive used drive space exceeds the RAID 5 limit.
