If you’re just running a few services, and will only ever be running a few services, I agree with you.

The additional burden of starting with proxmox (which is really just debian) is minimal and sets you up for the inevitable deluge of additional services you’ll end up wanting to run in a way that’s extensible and trivially snapshotable.

I was pretty bullish on “I don’t need a hypervisor” for a long time. I regret not jumping all-in on hypervisors earlier, regardless of the services I plan to run. Is the physical MACHINEs purpose to run services and be headless? Hypervisor. That is my conclusion as for what is the least work overall. I am very lazy.

Easily can have multiple LXCs, and being able to take snapshots for backup is probably a nice thing to have if you’re just learning.

And if they get more hardware, moving VMs to other clustered proxmox instances is a snap.

I don’t know much about static electricity and plastic, but would it be sufficient to ground off an arbitrary point in the case to the body of the PSU?

Why?

For the record, I fucking hate Plex.

But this is a disingenuous simplification of where the gap is.

Me, my brother-in-law, and friend all share our libraries with the same elderly relatives.

The GAP is that great grandma has to log in/out between servers to find content that may or may not be on an individual server. Plex lets you search/aggregate from all sources without having to jockey credentials and servers.

It’s not a giant ask. I heard a fucking absolutely brain-dead take that “that would require a centralized server which is against Jellyfins core ideology”.

So, I dunno. Maybe it isn’t YOUR use case, but it’s MY use case. Doesn’t make me a shill. I’m still pissed as hell.

But don’t fucking pretend that there is feature parity when there isn’t, and don’t accuse me of being a shill just because Jellyfin literally doesn’t support my use case. I WISH it did. I HATE PLEX.

Why are they proxying the stream through their server though

May or may not be applicable to your case, but often applications need additional configuration to work with a reverse proxy. Usually setting from what IPs it will accept forward headers from (your reverse proxy) and what the original requested host was (externally requested domain, eg: yourservice.yourdomain.com)

If your new setup has resulted in changes to either of those things, the issue might be a now-incorrect config of your apps behind the reverse proxy.

Oh nice, I’ll give that a shot. I was using IOTlink but the service wasn’t reliable on my machine and needed to be restarted constantly…

I’ll give HASS.agent a shot! Thanks

If you get a reliable way to sleep a windows machine via MQTT (not sure if that’s a route you’d take) but I’d be super interested in hearing about it.

I had a similar revelation. Home assistant has a WOL component, so you can set that up for easy starts. I’ve had mixed success with mechanisms to get HA to sleep the computer, though.

Ideally I want the machine to be sleeping I’d I’m not using it.

I’d never looked at them before, but yeah that super flower super modular supply looks pretty sweet. It looks like it has a ton of ports that I assume can be wired up as whatever you need.

For me, the splitters were just generic: they plug to an existing molex out connector and give you 5 SATAs on a ribbon.

https://a.co/d/gXtQ3Qp is what I’d bought, just for reference. The power supply I used them with wasn’t modular (ancient) and so whatever it had was what there was.

Maybe I misread, but if you are planning on having two different PSUs in play for the same system, it’s my understanding that it’s important to make sure the DC outputs share a common ground, which might be a little extra wiring.

Depending on how power hungry the drives are, and if your PSU has enough spare power, you can get cable splitters. I had some spare molex ports which I plugged a cable from Amazon that split it into 5 SATA power connectors.

You don’t want to infinitely split cables though, as tempting as that can be, because there are real electrical limits to doing that. Also just because a power supply is rated at X watts, that’s the total. Hard drives will use the 5V and 12V rails and usually there are individual limits on each rail.

Upgrading the PSU is another option. Probably the cleanest easiest best solution IMO. But even then, you probably can’t find a PSU that’ll give you 12 SATA connectors out of the box so you’ll probably need some splitters in there anyways.

In my case specifically, I’ve actually got a second power supply (because i already had it and it was otherwise just gathering dust) powering the extra drives. It’s a bit more complicated to get set up but, it’s an option as well.

Edit: also if you’re asking yourself where can you physically PUT the drives, I 3D printed these and slapped some fans on them:

https://www.thingiverse.com/thing:4875498

Yeah, that’s exactly what I’m talking about. I really like the idea of people with likewise sensibilities being able to act as their own cross-community mod team, without impacting other groups with different sensibilities.

An idea would be to allow “plug-in able” content sorting algorithms or content filters.

I hear so many stories of people slapping tons of filters in their clients (block all comments from ml users, block “Elon” and/or “Trump” keywords)… I think tons of people are running almost identical filters. Why not bake right into the Lemmy core the ability to pull filter sets from say, a public got repo?

Same with sorting. I’d love to have a “hot” algorithms that “punishes” posts based on comment sentiment analysis. Again, let me choose my sorting algorithm from a git repo. Let some person or persons develop a “good vibes” algorithm which keeps toxity off the top of my feed.

IMO, this is the way. Sorting by engagement has obvious issues. Introducing other weights to augment a system would make a huge difference in user experience.

You can’t change the people. Look at this comment section. OP said they don’t want to be yelled at and everyone took that cue to give a lecture. Completely no self awareness. Can’t change that.

But you can improve the algorithm. And IMO if you could crowdsource that dev in a way that doesn’t impose on mainline development.

As others have said, running out of motherboard SATA slots doesn’t mean you need a new machine to support expansion.

You can get m2 adapter slots for more SATA drives.

If you think you’ll be building a NAS in the future, and are cheap like I am, you might consider getting a pci-e expansion card for SAS rather than SATA drives. They’re backwards compatibile with SATA drives, but open you up to being able to use SAS drives which are common in enterprise data centers. You can get used lots of those drives on eBay WAY cheaper per TB when the data centers hour them out.

I’ve got a machine with 16 SAS drives running the unRaid OS, and I’m very happy with it for data hoarding and media serving. The drives (with shipping) cost $5/TB.

Whisper is fantastic and has different sized models so you can zero in to what gives you the best mix of speed/accuracy for whatever hardware you’ll be running it on

If the goal is to visualize growth trends, I don’t think raw user counts are the correct value to track on the Y-axis at all. That’s where my head was at when I said it doesn’t make sense. Abusing the Y-axis to try and coax data out in this case is just a symptom of having the wrong measure.

Daily new users. Percent user growth.

Yeah when I showed the cop the graph of my speed before getting in my car to be 67000mph (speed of the earth around the sun) to 67080mphwhen I was driving it he couldn’t see the difference so I didn’t get the ticket.

Or sometimes choosing a common-sense reference makes sense.

Which isn’t to say THIS one does, it doesn’t, but the absolutism of “it’s nerf or nothing” is a tad extreme.

It also assumes that you can’t directly sell power back to the grid, which without power efficient mining hardware would still be a more valuable thing to do with the electricity.

To follow on from the last commenter:

As the creator of a major subreddit, I can say that the collective masses can absolutely not be swayed from whatever the zeitgeist is… ESPECIALLY if the action is secret.

So, with that: since Lemmy is free and open source, I think bringing that discussion and possibly systemic solutions there (GitHub?) Would be the most practical way to affect tangible change.

Part of the solution might be to just accept that behaviour, and instead focous on how to mitigate the practical effects. User downvotes without corresponding comments maybe don’t effect the “net upvottedness”. Maybe sorting options that don’t include downvoting at all. If the concern is how naked downvotes affect visibility, maybe the resolution is in the visibility algorithms themselves.