Hello, my apoligize if questions already been posted
I would like to set up a new configuration with my Raspberry Pi 4 2GB and 2 hard drives of 3TB each.
I think to buy a SATA dock to plug 2x3"5 HDD, this dock is powered by a dedicated power adaptor and have 2 slots. What is your best recommendation to connect my two hard drives on raspberry since sata hat is sold out ?
If you have already tested drive bays or docks or enclosures with raspberry with USB let me know for good stuff
My first Node was an Rpi3+ directly connected on USB (singledrive).
It worked fine, response time and throughput on USB is good enough to run an PiNode
Later i decided to buy a NAS configured to Raid1 to prevent Nodeloss and to push up the resilience of the Network. If the Nas /Drivebay you choose is able to run docker directly you should be fine with it. If not like in my case you could try iSCSI. I have made the experience that iSCSI isn`t that bad as previously thought. I actually run 3 Nodes on iSCSI perfectly fine
Hello, thanks Solu for your answer but i dont want to use other computing unit than my Raspberry Pi 4, i have already check this with a cluster configuration that cost me more than revenue. Any other recommendation from you guys ?
I just spun up my second node, and both nodes are running on separate RPi 4s ( both are 4GB RAM version). They are low power solutions, which is super nice. Both nodes have a single HDD connected via USB 3.0.
From everything I’ve gathered here on the forum is that for almost all situations using a raspberry pi, the recommendation is to use a single drive per node. Although using your two drives to total ~6TB sounds like a good idea (more storage space, right?), the risk for one drive failing leading to total data loss is too great.
Here are a few of the threads that this has been discussed on:
Therefore, I believe what the Storj group would recommend, is use the existing hardware you currently have (RPi 4 (2GB) with one of your HDDs) to start your first node. Let that get going for a while, and then if/when your node becomes full with data you could then choose to start a second node, although this might require another computer/Rpi. I know there has been some discussion on running two nodes on the same Rpi, but I think the biggest bottle neck would be the Rpi itself, ranging from low power CPU, your version only having 2GB of RAM, and then also the USB controller trying to handle the I/O of the two HDD traffic. The reasoning behind the “staggered” timing is that two nodes behind the same public IP are “seen” as a single location and satellites would end up splitting traffic to both nodes, i.e. taking twice as long to finish the vetting process and loading the nodes up with data.
Here’s a couple threads explaining that second node process:
Hope this helps, and hope you choose to join the Storj network! I’ve had a good time setting up my node and learning about the project over the past 8 months. Plan to be here for the long haul.
Thank you very much for your answer, i bought a dual slots SATA dock and it’s ok, i have set up LVM with my 2x3TB disks. Raspberry pi can see two disks simultaneously. Thanks to the whole storj community.
Please, do not do that. With one disk failure the whole node is lost. It’s much better to run two separate nodes.
The only note - run the second node when the first one almost full, then the vetting process will not take forever.