Atemu

This is a false dichotomy. Just because containers make it easy to ship software, doesn't mean other means can't be equally easy.

NixOS achieves a greater ease of deployment than docker-compose and the like without any containers involved for instance.

I would not buy a CPU without seeing a real-world measurement of idle total system power consumption if you're concerned about energy (and therefore cost) efficiency in any way. Especially on desktop platforms where manufacturers historically do not care one bit about efficiency. You could easily spend many hundred € every year if it's bad. I was not able to find any measurements for that specific CPU.

Be faster at transcoding video. This is primarily so I can use PhotoPrism for video clips. Real-time transcoding 4K 80mbps video down to something streamabke would be nice. Despite getting QuickSync to work on the Celeron, I can’t pull more than 20fps unless I drop the output to like 640x480.

That shouldn't be the case. I'd look into getting this fixed properly before spending a ton of money for new hardware that you may not actually need. It smells like to me that encode or decode part aren't actually being done in hardware here.

What codec and pixel format are the source files?
How quickly can you decode them? Try running ffmpeg manually with VAAPI decode, -c copy, and a null sink on the files in question.

What codec are you trying to transcode to? Apollo lake can't encode HEVC 10 bit. Try encoding a testsrc (testsrc=duration=10:size=3840x2160:rate=30) to AVC 10 bit or HEVC 8 bit.

That doesn't matter. Snapshots only concern the state of any one filesystem; they do not address separate filesystems in any way.

That fully depends on where your kernels are stored. If they're in a separate partition, then no; it won't roll back with the rest of your system.

So you rolled back your root filesystem's system state to a snapshot but did you roll back the kernel you're booting into aswell?

If you didn't, that'd explain the symptoms seen here; you might be booting into a newer kernel than the system state has modules for. Without the appropriate kernel modules, Linux cannot mount filesystems or accept keyboard input. (This depends on which modules are required by the HW and whether they're built into the kernel or copied to the initrd though.)

For debugging, simply boot a live ISO and chroot into the system. The Arch Wiki has a page for that. You should be able to look at the journal inside the chroot and it'll tell you exactly what's wrong.

Note that ProtonMail and Fastmail have quite different feature sets.

ProtonMail does not store your Email in plain text for instance; they cannot read them or be ordered to read them. This comes with some drawbacks such as that standard protocols such as IMAP do not work without a bridge because they necessitates that the server can read all the emails.

I also find it weird that you can't create unlimited addresses on your custom domains.

For the shared domains, limits in this regard are absolutely understandable as the supply is limited but addresses should have next to no cost for PM when they're under my own domain.

Why is that? @protonmail@mastodon.social

Have you considered using Oracle's free VPS tier? Should be more than powerful enough to host a read-only Lemmy instance.

It's not ideal but if you're short on money, it's better than having your online data rot.

It really depends on what it is you're trying to share between machines.

I don't use syncthing but something that fulfils a similar function (git-annex). My Documents repo is set up in such a way that all instances of the repo try to have a copy of everything because documents are very important data and don't take much space. Other (larger) repos only try to have two or three independant copies; depending on how large and important their data is.

I would not "share" it synchronously as @gratux@lemmy.blahaj.zone recommended because in that case the data is only stored on one device and almost always accessed remotely. If the internet connection is gone, you'd no longer have access to the data and if the VPS dies, your data would be gone on all other machines too.

If you want to use Nextcloud anyways, that would be an option.

If all you want to do is have a shared synchronised state between multiple machines though, Syncthing would be a much lighter weight purpose-built alternative.

Since when does MS access run on IBM mainframes?

Thank @K900@lemmy.0upti.me ;)