One thing you probably need to figure out first: how are the dgpu and igpu connected to each other, and then which ports are connected to which gpu.

Everyone does funky shit with this, and you’ll sometimes have dgpus that require the igpu to do anything, or cases where the internal panel is only hooked up to the igpu (or only the dgpu), and the hdmi and display port and so on can be any damn thing.

So uh, before you get too deep in planning what gets which gpu, you probably need to see if the outputs you need support what you want to do.

It’s viable, but when you’re buying a DAS for the drives, figure out what the USB chipset is and make sure it’s not a flaky piece of crap.

Things have gotten better, but some random manufacturers are still using trash bridge chips and you’ll be in for a bad time. (By which I mean your drives will vanish in the middle of a write, and corrupt themselves.)

Am I missing something, or is this just the argo tunnel thing Cloudflare has offered for quite a while?

10000% this.

Tell me what it does, and SHOW me what it does.

Because guessing what the hell your thing looks like and behaves like is going to get me to bounce pretty much immediately because you’ve now made it where I have to figure out how to deploy your shit if I want to know. And, uh, generally, if you have no screenshots, you have no good documentation and thus it’s going to suuuuck.

It’s because of updates and who owns the support.

The postgres project makes the postgres container, the pict-rs project makes the pict-rs container, and so on.

When you make a monolithic container you’re now responsible for keeping your shit and everyone else’s updated, patched, and secured.

I don’t blame any dev for not wanting to own all that mess, and thus, you end up with seperate containers for each service.

I’d probably go with getting the ISP equipment into the dumbest mode possible, and putting your own router in it’s place, so option #2?

I know nothing about eero stuff, but can you maybe also put it into a mode that has it doing wifi-only, and no routing/bridging/whatever?

Then you can just leave the ISP router in place, and just use them for wifi (and probably turn off the wifi on the ISP router, while you’re in there).

Then the correct answer is ‘the one you won’t screw up’, honestly.

I’m a KISS proponent with security for most things, and uh, the more complicated it gets the more likely you are to either screw up unintentionally, or get annoyed at it, and do something dumb on purpose, even though you totally were going to fix it later.

Pick the one that makes sense, is easy for you to deploy and maintain, and won’t end up being so much of a hinderance you start making edge-case exceptions because those are the things that will 100% bite you in the ass later.

Seen so many people turn off a firewall or enable port forwarding or set a weak password or change permissions to something too permissive and just end up getting owned that have otherwise sane, if maybe over-complicated, security designs and do actually know what they’re doing, but just getting burned by wandering off from standards because what they implemented originally ends up being a pain to deal with in day-to-day use.

So yeah, figure out your concerns, figure out what you’re willing to tolerate in terms of inconvenience and maintenance, and then make sure you don’t ever deviate from there without stopping and taking a good look at what you’re doing, what could happen if you do it, and coming up with a worst-case scenario first.

What’s your concern here?

Like who are you envisioning trying to hack you, and why?

Because frankly, properly configured and permissioned (that is, stop using root for everything you run) container isolation is probably good enough for anything that’s not a nation state (barring some sort of issue with your container platform and it having an escape), and if it is a nation state you’re fucked anyways.

But more to your direct question: I actually use dns scopes and nginx acls to seperate public from private. I have a *.public and a *.private cname which points to either my external or internal IP, and ACLs in the nginx site configuration to scope where access is allowed.

You can’t access a *.private host outside the network, but can access either from inside it, and so (again, barring nginx having an oopsie somewhere) it’s reasonably secure and not accessible, and leaves a very clear set of logs (and I’m pulling those logs in and parsing them for anything suspicious and doing automated alerting if I find anything I would not otherwise expect) so I’m happy enough with the level of security that this is, when paired with the services built-in authentication options.

When you say you ‘can’t access local devices’ is it just via the browser, or can you also not ping/telnet/ssh/whatever?

If you don’t read the man pages, we’re going to ban you from the chat.

Better performance, but only for gay games, such as uh, well uh and um…

[Edit] I’ll have to see if I can find the video.

I can save you the time there, at least: https://youtu.be/hiwaxlttWow

There was a recent video from everyone’s favorite youtube Canadians that tested how many USB devices you can jam onto a single controller.

The takeaway they had was that modern AMD doesn’t seem to give a shit and will actually let you exceed the spec until it all crashes and dies, and Intel restricts it to where it’s guaranteed to work.

Different design philosophies, but as long as ‘might explode and die for no clear reason at some point once you have enough stuff connected’ is an acceptable outcome, AMD is the way to go.

I’m on year 5 with 6 of them and they’re all fine.

RTSP stream to frigate, and then frigate does the magic AI and recording shit.

They’re also not allowed outside the LAN and don’t seem to care about not being all internet connect-y, though YMMV on newer models.

I can’t think of a single case of being annoyed with them other than the mounting pressure is a little wonky and a sufficiently fat corvid can land on them and change the angle on one of the ones in the backyard but I’m not sure I’d blame the camera manufacturer because of a fat crow.

I’m more of a fan of :(){ :|:& };: personally.

(don’t run that unless you want to crash your shit)

I’d like to argue the other way: jump into whatever you find interesting and do stupid shit.

Unless this is a computer that your job relies on or whatever, then it doesn’t matter how much shit you break, as long as you’re learning what to not do next time.

It’s pretty much how everyone who is all ‘oh no! be very careful and take small steps!’ started, but they’ve just forgotten about having done that.

(I’ve broken so much shit both personally and professionally, but it just served to make me less stupid in the future. Or uh, at least unlikely to do the same stupid thing twice.)

generally replaced by systemd’s journald service

Basically this, and quite a long time ago. Anything even remotely modern (and by that I mean like, the last decade or so) is either using systemd, or in the case of debuntu, rsyslog.

Wonder what kind of funky environment is using syslog-ng, and to what scale so that there’s literally a ‘syslog-ng engineer’ job posting.

Yep, and there’s a shocking number of them to pick from: https://www.qemu.org/docs/master/system/qemu-cpu-models.html

So, this is a ~15 year old laptop?

The first two things that immediately come to mind when you’re kernel panicing is bad ram, and bad cpu temperatures.

Thermal paste doesn’t last forever, and it’s worth checking if your CPU or GPU are overheating, and repasting if so.

And, as always, a memtest is a quick and easy step to rule that out - I’d say half the “weird crashes” I’ve ever seen ends up being bad ram and well, at least it’s cheap and easy to replace?

Quicksync

Yeah, it doesn’t sound like you’re transcoding in a way that’ll show any particular benefit from Quicksync over AMF or anything else. My ‘it’s better’ use case would be something like streaming to a cell phone at 3-5mbps, and not something local or just making a file to save on your device.

DDR4 and no ECC

That’s what my build is: 128gb of Corsair whatever on a 10850k. I’m sure there’s been some silent corruption somewhere in some video file or whatever, but, honestly, I don’t care about the data enough to even bother with RAID, let alone ECC.

I will say, though, if you’re going to delve into something like ZFS, you should probably consider ECC since there are a lot more ‘well shits’ that can happen than what I’m doing (mergerfs + snapraid).

power consumption

A $30 or whatever they are kill-a-watt plus something like s-tui running on the NAS itself to watch what the CPU is doing in terms of power states and usage. I’ve got a 8-drive i9-10850k under 60w at “idle” which is not super low power, but it’s low enough that the cost of hardware to improve on it even a little bit (and it’d be a very little bit) has a ROI period of longer than I’d expect the hardware to last.