Nothing Special   »   [go: up one dir, main page]

Skip to content

pojntfx/ram-dl

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Project icon

ram-dl

A tool to download more RAM (yes, seriously!)


hydrun CI Matrix

Overview

ram-dl is a tool that allows you to upload and download RAM to/from your system. Well, sort of :P

It is mostly intended as a tech demo for r3map, a library for efficient distributed mmap with mounting and migration support, and go-nbd, a pure Go NBD server and client library. Despite this it can however also be of use for a few experimental use cases, such as:

  • Extending a system's memory capacity over the network: By leveraging the fast fRPC RPC framework, you can use ram-dl and ram-ul to expand your local system's RAM (effectively "downloading RAM") without having to use local disk space or memory.
  • Mounting a remote system's memory: By allocating a memory region with ram-ul (effectively "uploading RAM"), it is possible to mount a remote system's RAM to your local system with minimal latency.
  • Getting insights into memory usage and contents: By inspecting the "downloaded"/mounted memory with ram-ul, you can get insights into the contents of a remote system's memory.

ram-dl achieves this by essentially doing the following:

  1. Starting a go-nbd server and client locally with r3map's Direct Mount API
  2. Connecting the NBD server to a remote fRPC backend provided by ram-ul
  3. Calling mkswap, swapon and swapoff, which enables paging out to the block device provided by the NBD client and thus to the connected remote fRPC backend.

For most real-world use cases, using r3map's Managed Mount API or Migration API directly is the better (and much faster) option, but if you just want to see the Download More RAM meme implemented in real life, you've come to the right place!

Installation

Static binaries are available on GitHub releases.

On Linux, you can install them like so:

$ curl -L -o /tmp/ram-dl "https://github.com/pojntfx/ram-dl/releases/latest/download/ram-dl.linux-$(uname -m)"
$ curl -L -o /tmp/ram-ul "https://github.com/pojntfx/ram-dl/releases/latest/download/ram-ul.linux-$(uname -m)"
$ sudo install /tmp/ram-dl /usr/local/bin
$ sudo install /tmp/ram-ul /usr/local/bin

You can find binaries for more architectures on GitHub releases.

Tutorial

TL;DR: "Upload" RAM with ram-ul, "download" the RAM with ram-dl, done!

1. Upload RAM

On a remote (or local) system, first start ram-ul. This component exposes a memory region, file or directory as a fRPC server:

$ ram-ul --size 4294967296
2023/06/30 14:52:12 Listening on :1337

2. Download RAM

On your local system, start ram-dl. This will mount the remote system's exposed memory region, file or directory using fRPC and r3map as swap space, and umount it as soon as you interrupt the app:

$ sudo modprobe nbd
$ sudo ram-dl --raddr localhost:1337
2023/06/30 14:54:22 Connected to localhost:1337
2023/06/30 14:54:22 Ready on /dev/nbd0

This should give you an extra 4GB of local memory/swap space, without using up significant local memory (or disk space):

# Before
$ free -h
               total        used        free      shared  buff/cache   available
Mem:            30Gi       7.9Gi       6.5Gi       721Mi        16Gi        21Gi
Swap:          8.0Gi          0B       8.0Gi

# After
$ free -h
               total        used        free      shared  buff/cache   available
Mem:            30Gi       7.9Gi       6.5Gi       717Mi        16Gi        21Gi
Swap:           11Gi          0B        11Gi

🚀 That's it! We hope you have fun using ram-dl, and if you're interested in more like this, be sure to check out r3map!

Reference

ram-dl

$ ram-dl --help
Usage of ram-dl:
  -chunk-size int
    	Chunk size to use (default 4096)
  -chunking
    	Whether the backend requires to be interfaced with in fixed chunks (default true)
  -raddr string
    	Remote address for the fRPC r3map backend server (default "localhost:1337")
  -size int
    	Size of the memory region or file to allocate (default 4294967296)
  -verbose
    	Whether to enable verbose logging

ram-ul

$ ram-ul --help
Usage of ram-ul:
  -addr string
    	Listen address (default ":1337")
  -backend string
    	Backend to use (one of [file memory directory]) (default "file")
  -chunk-size int
    	Chunk size to use (default 4096)
  -chunking
    	Whether the backend requires to be interfaced with in fixed chunks in tests (default true)
  -location string
    	Backend's directory (for directory backend) or file (for file backend) (default "/tmp/ram-ul")
  -size int
    	Size of the memory region or file to allocate (default 4294967296)
  -verbose
    	Whether to enable verbose logging

Acknowledgements

Contributing

To contribute, please use the GitHub flow and follow our Code of Conduct.

To build and start a development version of ram-dl locally, run the following:

$ git clone https://github.com/pojntfx/ram-dl.git
$ cd ram-dl
$ make depend
$ make && sudo make install
$ sudo modprobe nbd
$ ram-ul
# In another terminal
$ sudo ram-dl

Have any questions or need help? Chat with us on Matrix!

License

ram-dl (c) 2024 Felicitas Pojtinger and contributors

SPDX-License-Identifier: AGPL-3.0