raphaelsc / blockv

Export block devices and access them using FUSE.

#BlockV

##Introduction

blockv project is composed of blockv server and blockv FUSE (client). blockv server is used to export a block device to the outside world. blockv FUSE is used to import a block device exported by a blockv server.

With blockv project, it's easy to mount locally a file system stored in a block device of a remote computer.

With blockv FUSE, it's easy to keep access to a large number of block devices that are exported by respective blockv servers.

This project is based on the concept of network block device. For further information: https://en.wikipedia.org/wiki/Network_block_device

##Getting started

Compile full project with:

g++ --std=c++14 `pkg-config fuse --cflags --libs` blockv_fuse.cc -o blockv_fuse;
g++ --std=c++14 blockv_server.cc -o blockv_server -lpthread;

There is no makefile because I am lazy, but I will write one as soon as possible.

##Playing with network-based block device At this section, you will learn how to export a file system to the outside world and import it from another machine.

Server side

1. Compile project for server side:

g++ --std=c++14 blockv_server.cc -o blockv_server -lpthread;

2. Create a pseudo block device that will be exported to the network:

qemu-img create -f raw -o size=500M ./pseudo_block_device.raw;

3. Format pseudo block device with a file system:

mkfs.ext2 ./pseudo_block_device.raw;
mkdir ./mount_point;
sudo mount -t ext2 ./pseudo_block_device.raw ./mount_point;
sudo umount ./mount_point;

NOTE: copy all files you want to share to the folder ./mount_point.

4. Export pseudo block device by running blockv server as follow:

./blockv_server ./pseudo_block_device.raw;

Read-only mode (write request is disallowed):

./blockv_server ./pseudo_block_device.raw --read-only;

Client side

1. Compile project for client side:

g++ --std=c++14 `pkg-config fuse --cflags --libs` blockv_fuse.cc -o blockv_fuse;

2. Mount blockv:

mkdir ./blockv_mount_point;
./blockv_fuse -d ./blockv_mount_point -o allow_root;

NOTE: *allow_root* option requires adding *user_allow_other* to */etc/fuse.conf*

3. Import the remote block device exported by blockv server:

ln -s localhost:22000 ./blockv_mount_point/remote_block_device;

NOTE: Replace localhost and 22000 by server ip and port, respectively.

It's possible to see the ip and port associated with a remote block device, look:

$ ls -l ./blockv_mount_point/remote_block_device;
lr--r--r--. 1 root root 524288000 Dec 31  1969 ./blockv_mount_point/remote_block_device -> localhost:22000

4. Mount the imported block device:

mkdir ./mount_point_for_remote_file_system;
sudo mount -t ext2 -o loop ./blockv_mount_point/remote_block_device ./mount_point_for_remote_file_system;

At this point, you can fully use the file system stored in the remote block device.

##Playing with memory-based block device

1. Mount blockv:

mkdir ./blockv_mount_point;
./blockv_fuse -d ./blockv_mount_point -o allow_root;

NOTE: *allow_root* option requires adding *user_allow_other* to */etc/fuse.conf*

2. Create a memory-based block device of 30MB:

truncate -s 30M ./blockv_mount_point/virtual_block_device;

3. Format the memory-based block device with ext2 file system:

mkfs.ext2 ./blockv_mount_point/virtual_block_device;

ext2 filesystem was chosen arbitrarily. Any other file system supported by Linux should work as fine.

4. Mount the ext2 filesystem with:

mkdir ./mount_point2;
sudo mount -t ext2 -o loop ./blockv_mount_point/virtual_block_device ./mount_point2;

At this point, you can fully use the file system stored in the memory-based block device.