“I currently decided to store the snapshot on the same hard drive of my installation”
That wont do you any good if you get a harddrive crash or need to reinstall. The way i would do it is to make snapshots on a USB and also keep another liveISO USB. That way i can reinstall, then just use timeshift to restore.
To second what others have said: VM’s aren’t suitable for gaming regrettably.
PopOS is a rather reliable distro, and I personally have loved the window tiling features they added, but it should be noted that they only have LTS from a year ago at the moment. I think that’s just while they work on their new desktop environment, but the older packages might be a tad bit of a transition coming from Arch.
Are you going to add support for the new mobos? I’m not :P it’s the open source dream. I think manufacturers should contribute to coreboot and use it, and probably will eventually. Think of the savings in licensing cost :P plus they get free labour from the community sometimes.
People there just blindly hate nvidia and praise amd. And when I tried once tell the fact, that amd opensource drivers suck - I got a lot of minuses at my comment. Oh well, anyway, I really hope, that all that people will buy amd card one day and suffer as I did. I doubt that will change their opinion, cause they will still hate just_working nvidia drivers, cause “oh no, they not support VRR on my experimental wayland DE”. They focking dumb man, I tell ya. Anyway, I will try to sell my “awesome and opensource” amd 7900xtx and buy cheaper nvidia card. Just cause at cheaper nvidia card I could at least play fucking games and it will not crash my video driver every now and then.
Ditto. I use the Nvidia driver in Arch, and I forgot that I had it installed. No problems on this machine for more than 6y.
I have never heard of an installer running in the background causing weird delays requiring weird routines. Maybe it’s the way it’s packaged on their distro, perhaps? Like they’re shipping the installer in the package instead of the contents of the installer?
O use ext4 at home and in servers that are not SLES HANA DB ones.
On SLES HANA servers I use ext4 for everything but the database partitions, for which SAP and SUSE support and recommend XFS.
In a few occasions people left the non-db partitions as the default on SUSE install, btrfs, with default settings. That turned out to cause unnecessary disk and processor usage.
I would be ashamed of justifying btrfs on a server for the possibility of undoing “broken things”. Maybe in a distro hopping, system tinkering, unstable release home computer, but not in a server. You don’t play around in a server to “break things” that often. Linux (differently from Windows) servers don’t break themselves at the software level. For hardware breakages, there’s RAID, backups, and HA reduntant systems, because if it’s a hardware issue btrfs isn’t going to save you - even if you get back that corrupted file, you won’t keep running in that hardware, nor trust that “this” was the only and last file it corrupted.
EDIT: somewhat offtopic: I never use LVM. Call me paranoid and old fashioned, but I really prefer knowing where my data is, whole.
Most comments suggesting btrfs were justifying it for the possibility of rolling back to a previous state of files when something breaks (not a btrfs breakage, but mishaps on the system requiring an “undo”).
Ah, I see. While that use may be a good plan for home server, doing that for production server seems like a bandaid solution to having a test server and controlling deployed changes very carefully.
It really depends on your priorities. Single drive is good for a home system with nothing really important on it… once you get to wanting to keep it and where recovery from backups is too much downtime, you want at least a drive mirror… nothing wrong with exr4+mdraid for that, although you don’t get the checksumming that zfs gives it will be pretty fast & if a drive fails you can run degraded on one drive until you get the new drive in.
I’ve been running zfs for 10 years and not lost a single byte of data even after doing stupid shit like tripping over the sata cables and disconnecting half the drives. It’s survived multiple drive failures (as long as the failures are on different bits of the disk, it recover get a clean copy onto a third drive, but it’s a brown trousers moment when stuff like that happens).
Downsides, it aint fast, and it does tend to like lots of memory. You want it on your fileserver, not your gaming system.
IMO there’s no point in a single drive zfs… it’ll warn you faster that the drive is f*cked but what do you do then?
I agree. Love ZFS for the NAS, but for a single drive desktop system, it is almost pointless and in my experience slower for desktop usage. ZFS is great for what it was designed for.
Ext4 is old, but fast and very robust. You won’t loose data or corrupt the filesystem if your system looses power. It can even survive partial wipes, if you accidentally overwrite the first few megs of you drive with a messed up dd, nearly all your data will be recoverable, including filenames and directory structure.
It doesn’t have very fancy features, but it is the best tested and most robust option available. (also the fastest due to its simplicity)
Btrfs has things like copy on write files that can protect you from an accidental rm, but this won’t save you from drive failures, so you still need backups for important data.
You won’t loose data or corrupt the filesystem if your system looses power.
Some secondary storage devices ignore standards and outright lie about sectors being successfully written when they are actually scheduled to be written out of order. This causes obvious problems when power failure prevents the true writes from completing. Nothing can be guaranteed for such drives.
linux
Hot
This magazine is from a federated server and may be incomplete. Browse more on the original instance.