Okay, I just stumbled across LBB. Nice project. However, coming here as a Systemadministrator I have a few thoughts:
a)
I know the SD card is mounted and an ID is written to the card. However, I would prefer the card is remounted ReadOnly after the ID is written. That way the content of the card should be more secure in case of any unplanned shutdowns or power cuts.
b)
I had some problem with the /dev names of the external storage as well as the SD card. I do have a USB-C hub with a card reader. If I connect the external HDD to the USB3 port of the HUB it’s seen as /dev/sdc while when connecting it directly to the Raspberry USB3 port it’s seen as /dev/sda. So
So the questions is: Is there an easy way to identify if the attached storage device is either a SSD/HDD or an SD card connected via a card reader? And if not, maybe just go via the size of the storage. So the “small” one is always the card while the “large” one is always the storage device?
c)
Are there any plans for further improvements? I’m thinking about building proxy media. Starting to record more videos lately. It would be great to use ffmpeg and generate smaller size videos for previewing.
I was always looking towards the Gnarbox, but I think the hardware is kind of expensive to what it delivers. A Raspberry Pi 4 should be powerful enough to do the same work.
I know the SD card is mounted and an ID is written to the card. However, I would prefer the card is remounted ReadOnly after the ID is written. That way the content of the card should be more secure in case of any unplanned shutdowns or power cuts.
Good point. I’ll see what I can do.
So the questions is: Is there an easy way to identify if the attached storage device is either a SSD/HDD or an SD card connected via a card reader?
I have been grappling with this for almost as long as the project exists. So far I wasn’t able to come up with a better alternative to the current implementation.
So the “small” one is always the card while the “large” one is always the storage device?
That doesn’t strike me as a robust solution. Sorry!
Are there any plans for further improvements?
Plans are mostly determined by my own needs and by user requests. If you have an idea, file a issue on GitHub, describing your idea.
And another thought: Why not build the system on a read-only linux partition. Like other embedded linux solutions do. That way an unclean shutdown or power cut will not giving any risk to the OS corruption itself.
I know I know Just a quick idea…it’s just so strange that the device names change depending on where I insert the drives. So I need to find a solution…
Also, the idea is that the end user can easily install Little Backup Box on the existing system. Setting up a read-only system would require some manual work, and I’d like to avoid that. Little Backup Box is aimed at photographers and people without Linux skills after all.
To answer to your question b, you can set mount rules according tu the UUID of the file sysem. This property is specific to the medium so it won’t matter if udev or whatever daemon is attributing it sda or sdc …
You can manage this settings in /etc/fstab or via udev rules
Sure, there are plenty of ways to mount a storage device. The problem is how do you mount any device without a) any Linux skills and b) editing anything (e.g., /etc/fstab) manually.
Think of this from the end-user point of view. What you want is for the system to recognize any storage device you plug into it, with zero interaction. You are not a Linux expert, and you don’t want to edit or configure anything. The current implementation is not perfect but it does address these two points.
So the perfect solution would be to provide a complete image
That’s not my intention. I don’t have resources and time to maintain an image. And again, I prefer to give users the ability to install Little Backup Box on an existing system.
Especially with no Linux Skills, it would be difficult to explain people that they should not unplug the power to shutdown the system.
Lack of Linux skills doesn’t imply the lack of common sense. If you shut down a machine by unplugging it, maybe you shouldn’t be using Little Backup Box (or any computer for that matter) in the first place.
I think it would be possible to create a read-only Linux via a script without manual interaction.
If you have an idea, file a issue on GitHub, describing your idea.
Yes, I don’t think that this approach is fool-proof.