Hello, I am building a new storage device with 36 HDDs and looking for some direction on using FreeNAS/ZFS. My goal is to add HDDs overtime with very minimal changes to the system. When I pop in a new HDD, the system identifies space and the plotted starts to plot and push K32 files until it fills in the space. Let me know if anyone using ZFS and what is the configuration.
Is your plan to use the entire storage device for Chia farming? If so, ZFS might be a bit overkill, unless you are specifically looking to do something with ZFS. It has generally been recommended that you trade parity space for more farming space.
good point. Will scrap the ZFS idea.
I’m using TrueNAS to plot, but not to farm. I’m only using it because I already have it running and had extra CPU/SSD/RAM to spare for plotting.
i have a truenas to store my plots för farming, its a datastore for my esxi farm so one vm have a disk where the plots are, but im thinking of expanding the truenas with som jbod if the price of chia is okey so it is worth it, i still need more space for other things so i will upgrade anyway
As one of the TrueNAS devs, I’m using TrueNAS SCALE 21.06 exclusively for farming and plotting. Plots are stored on ZFS pools, one disk == one pool. No redundancy required, but still get all the zfs caching and checksum goodness.
Further benefit is setting up plot scratch space on ZFS. You can stripe hdds to get SSD speeds or stripe SSDs to get even more performance
We are provided a custom docker image for TrueNAS (and non-TrueNAS users) to make it even more convenient.
I am not into ZFS or any raid with drives on my farm yet. However, just asking, with stripe HDDs or raid 0, if any one of the HDDs breaks, all your plots that’s striped across all HDDs would be invalid, thus needed to be replaced with new plots, right?
That’s the reason why I’m not going into raid yet.
Yes, I run with no redundancy on the farming drives. Plots are disposable, old plots have no more chance of winning than a new plot. So if/when a drive dies, I’ll just replace with a new drive and replot.
Just to clarify, I only use stripes for plotting temp space, since it’s all about speed.
hey! thanks for posting here and for the work you do w/ TrueNAS. What’s the real difference for Chia farming purposes between CORE and SCALE? thx
CORE works great if you just plan to export the plots over NFS/SMB to some other harvester / full node. Running Chia software on it has proven difficult, since few folks run Chia on FreeBSD and it’s not well supported.
SCALE on the other hand can easily do that, as well as running the full node / harvester directly on the storage. Thats how I do my farm, with each SCALE system running a remote harvester only, connected to my full node on a more powerful box.