testing the distributed part of dist-fs

[akielaries]

The main storage I am working with is a 500gb Seagate RSS LLC Slim SSD on the Jetson Orin Nano in the home lab. Plug in a small 8gb thumbdrive into one of the raspberry pis/some other device in the rack to try sending packets over a familiar socket and writing to that thumbdrive "remotely"

[akielaries]

There should actually be 2 options for the "distributed" part of this.

1. network based interface (NFS like implementation)

• one hosts the drive "mount", this will be called the host
• N number of clients that can connect to the host
• this is very sysd-esque, particularly the clients getting data from the host
• since I am testing with single board computers (Raspberry Pis and Jetson Nanos), I'm not very familiar with the use

2. physical interface

• SPI would be our best bet for serial implementation as it has the highest transfer speed
• Adding I2C will be nice after
• STARTING with UART would be nice to get something flowing
• SPI will at least allow us to do some daisy chaining

OPTION 2: this will entail a daemon (constant running process) on the host constantly polling for commands defined in packet.h. these commands will contain information like what operation to run (list, upload, delete, download), additional info like name of the file. the client will be putting this commands together via the dist-fs.cpp CLI file with a combination of what is read from the dist_fs.conf file. dist_fs.conf will determine how we communicate with the "drive" on the host. options would be directly (running on the host) and the "distributed" part, over a physical connection (UART, SPI, I2C), or over the network(TODO)

[akielaries]

UART is incredibly slow, time to get SPI and network comms working

[akielaries]

SPI is a pain to configure especially on the fly. trying network based comms next

[akielaries]

this should be more promising just need to get it working in the comms interface