Ok, some users reported bad arch mirrors when using netinstall, we’ll need to investigate that further. Don’t worry about it now, i’m just collecting information…
Thank you very much for answering!
I had a lot of errors with nouveau and wayland, I had to disable nouveau (blacklist)
now the pc is working fine again just with the integrated intel video card 
I suppose that I can not use nvidia proprietary driver, since I’ve read that it’s not working with wayland
Both btrfs and lvm snapshots are stored on the same partition you’re making a snapshot of, so the snapshots usually won’t be of any use to you in case of ‘partition damage’.
It is possible to copy and restore snapshots to/from other storage locations (for example with btrfs send/receive or dd for lvm) but this involves moving a lot of data around. You’re probably better off using something like clonezilla, which will also restore the disk layout if needed, if you need a disaster proof backup.
As I use and need my Linux system for work also, my use case for snapshots is primarily to quickly recover from my own stupidity or a botched upgrade. I could probably fix most arch update issues but sometimes lack the time and/or motivation. Most update issues get resolved upstream quickly so I’ll mostly just restore and try again somewhat later.
BTRFS snapshots (go for it!)
If you really want to make use of snapshots I would deeply recommend you try out btrfs. If you follow my BTRFSonLUKS guide you’ll have a system in which even a total ignorant user should be able to create and restore snapshots with a GUI (Timeshift).
LVM snapshots
If you want to use LVM snapshots things get a lot more complicated.
Most LVM setup guides, including my LVMonLUKS tutorial, assign all the available space of the volume group to the logical volumes (root, home, maybe also swap). If you run sudo vgdisplay you’ll notice there’s no free space in the volume group (‘Free PE / Size’ = 0).
The problem: LVM snapshots are created and stored in the FREE space of a volume group. You will first have to free up space by reducing the size of one or more logical volumes (home or root for example). For example, if you want to free up 30GB you will have to do this from a different environment (e. g. EndeavourOS ISO) by unlocking your luks partition, then shrinking the filesystem (mostly ext4) by more than 30GB, then shrinking the logical volume by 30GB, followed by growing the file system to its maximum. You could now store up to 30GB worth of snapshots.
Unlike btrfs snapshots you allocate a fixed amount of space, let’s assume 5GB, when creating a LVM snapshot. Your snapshot will become corrupted and unusable if more data than 5GB changes on your system. So you have to keep an eye on that. There is a way to dynamically enlarge a snapshot if needed but this isn’t the default configuration.
Unlike btrfs snapshots, to my knowledge, there is no way of creating, restoring and managing lvm snapshots with the help of a GUI. You will have to use terminal commands.
Unlike btrfs snapshots, lvm snapshots are ‘destroyed’ when restored (‘merged’) to that state. This merging process can take quite some time for large snapshots; btrfs snapshots are merely moved which is nearly instantaneous.
Don’t get me wrong, I personally also still use LVM-snapshots. They have never let me down but I’m gradually converting my systems to btrfs because it’s much less work and time consuming restoring snapshots.
Of course, if you still want to go the LVM route against my advice, I’d be happy to provide the basic commands to get you started.
3 Likes
First: Thanks a lot for all your suggestions and your time! 
As I do not have space left in the LVM drives on all PCs with LVMonLUKS I would have to reinstall every PC. Especially the installation of the family “server” took more then 4 weeks because of all the things I had to learn about the differences between Ubuntu and Arch. I do not think my family would accept, that the family “server” is down for more then some hours … 
For that case I am using two things:
etckeeperfor restoring my misconfiguration.- I do updates on one of the PCs/laptops I do not need most of the time. If there is a problem, I wait with the other PCs until the problem is solved.
If I will install my next PC for the family, I will try to follow your tutorial.
Interesting concept; will look into it, thanks. I’d probably combine this with snapshots though.
I want as little to do with the administration of my family members PC’s - the ease of working with btrfs snapshots has severely reduced the need for me to get involved. 
I try to do this with my three active Arch systems too, but if you wait too long there could then be a totally different problem arising … 
see you
1 Like
Small issue I just noticed, when you get to the part of the installer where there is a check box for using strong passwords it is unchecked by default but the installer still requires a strong password.
1 Like
Feel free to ask, I am using that since the days I started with Ubuntu more then 10 years ago. A friend of mine gave me that hint after I completely “destroyed” one of the family PCs.
I am using openLDAP and NFS for the family PCs. All the user data is stored on the family “server”. The clients only have a local admin user and the programs. Setting up a client needs about two hours, including things like printing, UPS configuration etc.
I check updates daily or more often. If install is fine, I do it with all clients, except the family “server”. There I install the kernel updates only once a week.
1 Like
I’m not sure if I understood you correctly, but I think you solved this in #5 of you 1st tutorial. Once the steps have been taken there’s a snapshot space available, as sudo vgdisplay shows at the end… For example, in my case, it was 24.4 GB.
You’re right! How forward thinking of me.
To be honest, I totally forgot about that and assumed I’d assigned all the space.
@2000
In your great tutorial No. 1 (LVMonLUKS) on the our wiki, in #13 there’s a line starting with:
"cryptdevice=UUID=..."
I’m not sure but it seems like should be placed in the line:
GRUB_CMDLINE_LINUX
From the Archwiki:
To pass custom additional arguments to the Linux image, you can set the
GRUB_CMDLINE_LINUX+GRUB_CMDLINE_LINUX_DEFAULTvariables in/etc/default/grub. The two are appended to each other and passed to kernel when generating regular boot entries. For the recovery boot entry, onlyGRUB_CMDLINE_LINUXis used in the generation.It is not necessary to use both, but can be useful.
In short …
- Options in
GRUB_CMDLINE_LINUXare always effective. - Options in
GRUB_CMDLINE_LINUX_DEFAULTare effective ONLY during normal boot (NOT during recovery mode).
So, yes, technically you’re right, it would be better to place "cryptdevice=UUID=..." in GRUB_CMDLINE_LINUX!
But, …
seeing that EndeavourOS seems to set GRUB_DISABLE_RECOVERY=true as a default, the GRUB recovery entries aren’t generated anyway, so it won’t matter to the average user and I personally don’t bother also.
I still use this method on one laptop and has been working fine for about 14 months 
. Maybe BTRFSonLUKS seems like a better solution and has some advantages (works on the second machine), but on the other hand, he didn’t get up just once, after some update, so I continue with this…
Can you write please if minor changes within /etc/default/grub between the first and second tutorial versions have a similar meaning as above (except ‘quiet’)? I mean adding this line:
root=/dev/mapper/vg0-lvroot
In other words, would it be advisable to add this line?
Yes, just as cryptdevice=, root=... should actually better be placed in GRUB_CMDLINE_LINUX.
1 Like
OK, thanks.
Greetings!
where is this set?
It’s preset in /etc/default/grub; I certainly didn’t add it manually …