I have created a new sectioned volume in windows 11 disk manager.
I have used a 500GB HDD and a 128GB SSD. They make a total of 628Gb.
I have started copying the files with robocopy to the new volume. The files seem to be being copied.
When I look at the new 628Gb volume it appears to be empty. It has 150Mb occupied and does not seem to be filling up.
It has been copying files for a few hours now.
It is normal? Is the sectioned volume valid for storj?
Robocopy has already finished copying.
Nothing was copied.
What will be the problem?
This is a very bad idea to use striped/simple/JBOD/mergefs/RAID0 volumes - with one disk failure the whole node is lost. And you already have problems with copying files.
I would not suggest to use striped volumes (in Windows Storage Spaces it’s called “Simple”) like this. You may use RAID if you want, but it should be either mirrored (“Two-way mirror” or “Three-way mirror”) or with redundancy (“Parity”).
However, it’s a bad idea to combine spinning drives with SSD in a one storage pool like you did, usually SSD configured in the pool as a cache:
Joining two units is to take advantage of the low capacity units I have. I have tried to sell them but buyers are not interested in them.
I’m going to try with windows storage space.
It wouldn’t help. To have a normal RAID with parity you need to have at least 3 disks, but since you have only two, it’s not an option.
The mirror can be created only in size of the smallest disk, so no luck too.
You can run a node on HDD, SSD can be used for OS. However, if you already have a SSD for your OS, then perhaps the only option is this RAID0, but it’s risky.
I agree that creating simple raid won’t be productive and btw raid0 won’t work either, disks shall be the same size. JBOD will work but doing so will be wasting an opportunity to do much better.
Instead, the situation is ripe for using block level tiered storage: metadata would end up on SSD and data stay on hdd, accelerating lookups and shortening node response time.
There are a bunch of solution on the market, but I’m not up to date on windows to recommend specific ones.
From memory, and a bit searching:
why x2 the risk of losing the node data? =.= buy a disk man, it’s cheap. :v
Because everything has monetary value. Buying a disk is $100. Losing a node at the double risk — (risk of drive failure)*2*(worth of node). Ballparking worst numbers, 0.02*2*$50 = $2.
Obviously, buying a disk is 50x worse decision than using tiered storage.
$100 you can get a 4TB disk, even if you don’t run Storj anymore you can use it for other purposes. Lol.
$100 can get you $12TB enterprise disk but that’s beyond the point.
I too agree that that hardware must be recycled as it is obsolete. But if OP had need for disk for other reasons they would have already bought one. It seems they want to find use for existing hardware.
I understand your argument.
500GB HDD and 128 SSD… seems to me some retired laptops. So 5400 rpm laptop HDD is crap anyway. You better buy a new one and use SSD for cache. You don’t get any money from 500 GB anyway.
Just like electricity needed to power the drive. Assuming a drive takes 5W and electricity price is, let say, 0.36 USD/kWh, you need ~800 GiB of data to pay off just the electricity it takes.
Electricity here is 0.20usd Kwh, with taxes included.
I’m trying to reuse units that I don’t use.
It is not known when the units will fail. They may fail after many years of use.
So you have three options in this case:
- use a tiered Storage Spaces
- use it as JBOD (Simple volume) in Storage Spaces
- JBOD created from the Disk Manager (but this option seems doesn’t work in your case).