I have an unraid server that is going to be repurposed and looking at being a storj node operator. I’m looking for another node operators input and advice.
How many threads available in the CPU’s? if you have ~20 or more you might want to look at golem.
unraid has the disadvantage of not having iSCSI block based storage services available. I’m not sure if there is a optimal way to use it with unraid besides visualizing another service and passing through some physical drives for it. Something like FreeNAS, TrueNAS would get you a high performing storage system.
I fully plan on reformatting probably leaning towards an Alpine Linux install with zfs
Dual Xeon 2670
1tb nmve cache
1 rtx 4000
and a supermicro 18 bay drive.
Hello @d1r7y and welcome to the forum!
You may repurpose a few TB for Storj, but it wouldn’t make sense to allocate 100TB for it. That’s way too much space that would never ever get used for the coming years in all likelihood.
Have a look at this community made estimator to get a better view on what to expect (storage and ROI wise):
There is one exception though: If you happen to have several IPs in different /24 subnets, then you may consider allocating more TB to Storj, scattered among several nodes. But that’s a pretty rare circumstance.
Would not work. It’s better to use FreeBSD or Ubuntu.
Proxmox is ok you could give truenas a try and get more out of your server both of which use zfs.
Not sure if it is worth going this route anymore, but I know people that are splitting large disk pools between Storj sharing and Chia plots. Allocate some space for Storj and fill the rest with Chia. As your Storj fills up expand that space and cut back the Chia.
Any there any tasks being paid for on golem-network?
That’s what I do. My available space is spread about 50/50 over Storj and Chia. Chia barely makes me any money, but it’s better than nothing. Currently about 19TB each.
Do you solo farm or mine in a pool?
Pool, otherwise it would still take me between 1 and 2 years to get a block.
I think I actually found my first block solo with less than 19TB shared…and I have been quite late to the party because the crappy client didn’t sync after ISP reconnects…so imho soloing is still an option if you’re feeling lucky
Yes, but you need a high number of cores/threads to get work.
(Which stands to reason - other wise what is the point of offloading the work.)
I’d say you need at least above 20 threads. if you look at the stats about 5% of the computers on line have tasks at the moment.
I’m thinking for myself to work towards picking up a Dell t420 dual socket machine. With the top CPU for that hardware I can have 20 cores/40 threads for not such a bad price.
I found my first block with around 1-2TB shared in the first week I started. The estimate to win was 3 months at the time, so I got very lucky as well… but yeah, I’m not going to count on luck alone if I don’t have to. Even with pool mining finding a block still gets you 0.25XCH though. So it’s still nice if it happens, I’m just not counting on it. I did unfortunately have to replot when I made the switch.
Are you using SSD’s at all? How is the wear going if so?
I’m still in the process of replotting, good thing I had not even filled all my space before pooling went official.
No, I intentionally created a non-SSD accelerated volume on my NAS. I don’t expand often, so I’m fine with slow plotting when I do. Not worth the wear and tear. I did use SSDs for a while, but I churned through half the lifetime of one of my SSDs. The economics didn’t add up. So I’m just taking my time now.
In my case I’m even plotting one at a time. But if you have multiple HDD’s that aren’t in an array, you could plot one per HDD and it would still be fairly quick.
I have way above 1 PB written on my 980 Pro 1TB boot drive, still working fine. Warranty runs out at 600TBW on this model.
100TB is going to take forever to fill up. I’ve allocated 8TB and am nowhere close to that after about 6 months.