I just upgraded to a Xeon E5 v4 processor.

I think the max RAM on it is about 1.5 TiB per processor or something.

It’s not new, but it’s not that old either. Still cost me a pretty penny.

One thing that was recommended to me by someone a while ago, is that, unless you need it for something specific, mount your media in Plex as read only.

Plex has functions where you can delete content from the library from their UI. If you need that for some reason, obviously don’t make it read only. If you’re hoarding the data, and therefore never delete it, or use an external system for deleting files, then RO all the way.

The only caveat to this is if you’re using a local disk on the Plex system, which then shares out the drive/folder for adding new content, in which case, you’re screwed. It has to be rw so the OS can add/remove data.

In my case, as I think may be common (or at least, not rare), my back end data for Plex Media is on a NAS, so it’s easy to simply have the system running Plex, mount that network share as RO, and you’re done. The data on the NAS can be accessed and managed by other systems RW, direct to the NAS.

Since Plex is exposed to the internet, if anyone with sufficient rights is compromised, in theory, an attacker could delete the entire contents of your media folder with it. If you limit RW access to internal systems only, then that risk can be effectively mitigated.

Depends on the UPS. Many cheap offline UPS units don’t. Anything line interactive or online will.

APC makes low end offline UPS units, which are cheap garbage.

They also make line interactive and online ups units, which are decidedly not completely garbage.

I pick up line interactive APC units from used locations like eBay, and go buy off label replacement batteries. Haven’t had any problems with them so far.

To date, over the last ~10 years of running a homelab, I have used mainly SMT 1500 units, one was a rack mount. I’ve recently upgraded to an SMX2000. I’ve replaced batteries, but never a UPS, and never any server components due to power issues. I’ve run servers ranging from a Dell PE 2950, to a full c6100 chassis, plus several networking devices, including firewalls, routers and PoE switches. Not a single power related issue with any of them.

I’m just happy that Lemmy.ca made the list.

Yay centralization!

I moved my DNS to a pair of raspberry Pi 3’s running bind, with a DNS stub zone for my homelab domain that points to my homelab DNS servers.

That way the internet keeps working whether my homelab works or not. Keeps the wife aggro down.

You can do whatever you want. Don’t let anyone tell you it’s “wrong”. A big part of homelabbing is to try stuff. If it doesn’t work, that’s fine, you learned something, and that was the point.

For me, I don’t see a UPS as essential. It’s generally a good idea, but not strictly essential. My servers are on 24/7, because I have services that do things overnight for me. I also know that some people access my lab when I’m not awake, so I just leave it on so it can be ready for anything at any time. It poses some unique challenges sometimes when running stuff that’s basically 24/7/365.

Be safe, have fun, learn stuff.

Yep, I’m sure they do.

Realistically, does any average consumer know what’s on which circuit?

Spanning the split phase will screw you up, across breakers won’t be fun but shouldn’t pose any serious problems, as long as it’s not in different sides of the split phase.

I’m pretty sure they say this because actually explaining what will work and what won’t either requires significant prior knowledge of power systems, or a couple of paragraphs of explainers before you can get a rough picture of what the hell they’re driving at.

Everyone I know who has used powerline, just plug it in and see if it works. Those who were lucky, say it’s great and works without issue, etc. Those who were not lucky say the opposite.

I’m just over here watching the fireworks, eating popcorn.

I’ve been doing IT work for more than a decade, I was a nerd/“computer guy” well before that. I’ve had a focus on networking in the past 15-20 years. You learn a few things.

I try to be humble and learn what I can where I can, I know that I definitely do not know everything about it, and at the same time I try to be generous and share what I’ve learned when I can.

So if you have questions, just ask. I either already know, or I can at least point you in the right direction.

It definitely sounds like you have some challenges ahead. I personally prefer MoCA over wireless, simply because you can control what devices are able to be a part of the network, and reduce the overall interference from external sources and connections.

With WiFi, being half duplex, only one station can transmit at a time (with come caveats). Whether that station is a part of your network, or it is simply operating on the same frequency/channel, doesn’t matter. So in high density environments, you can kind of get screwed by neighbors.

MoCA is also half duplex (at least it was the last time I checked) so having a 2.5G MoCA link, to a 1GbE connection (on the ethernet side) should provide similar, or the same experience as pure ethernet (1G full duplex)… The “extra” bandwidth on the MoCA will allow for each station to send and receive at approximately 1Gbps without stepping on eachother so much that you have degraded performance.

However, it really depends on your situation to say what should or shouldn’t be setup. I don’t know your bandwidth requirements, so I can’t really say. The nice thing about ethernet is that it on switched networks (which is what you’ll be using for gigabit), the. Ethernet kind of naturally defaults to the shortest path, unless you’re doing something foolish with it (like intentionally messing with STP to push traffic in a particular direction). The issue with that is that ethernet doesn’t really scale beyond a few thousand nodes. Not an issue for even a fairly large LAN, but that’s the reason we don’t use it for internet (wan side) traffic routing. But now I’m off topic.

Given the naturally shortest-path behavior of ethernet, of you have a switch in your office and you only really use your NAS from your office PC, you’ll have a full speed experience. If nothing else needs high-speed access to the NAS, you’ll be fine.

Apart from the NAS or any other LAN resources, the network should be sufficient to fully saturate your internet connection. So the average WiFi speeds should be targeted towards something faster than your internet link (again, half duplex factors in here). I don’t know your internet speed so I’m not going to even guess what the numbers should be, but I personally aim for double my internet speed for maximum throughput on my WiFi as much as I can. The closer you can get to doubling your internet speed here, the better. Anything more than that will likely be wasted.

There’s a ton to say about WiFi and performance optimization, but I’ll leave it alone unless you ask about it further.

Good luck.

It can be faster, it really depends on whether you have a clear-ish channel for the mesh, which is why I would recommend something on the higher end, hopefully with a dedicated radio for mesh, so it can be on a different channel with (hopefully) less interference.

If the mesh radio is shared with client access, or if it’s on a busy channel, it may be much, much slower than some options.

Hello friend.

You can get 2.5gbps MoCA now. You may want to consider upgrading.

Nice username btw

Depending on where you live and what your power circuits look like (not the outlets, the circuits that power them), you may have a great, or very poor experience.

I’d need to know what country you live in to know more, since power wiring standards vary from country to country. In the USA and Canada (I’m in Canada and the USA is the same), we use split phase and crossing the split phase will severely hinder the ability for powerline to perform.

It’s a viable option, not my favorite option, I’d recommend MoCA (coax) over powerline, but it’s ultimately up to you.

IMO, powerline is going to depend on a lot of factors including what kind of power you use, which varies from country to country. Where I am in North America, we use 240v split phase, and the powerline adapters are 120v (half phase), so if one unit ends up on one side of the phase, and one ends up on the other side of the phase, you’re going to have a bad time, if it links at all… So knowing which “side” of the split phase your powerline is on becomes critical, which is not something most people know about their power situation. As a result, it’s basically a crap shoot whether it will work well or not.

I have three suggestions for you.

Easy mode: find a triple radio mesh wifi system and get at least two nodes. Generally the LAN Jack on the satellite nodes will bridge to the LAN over WiFi. Just add a switch and use it normally. This will harm your overall speeds when connecting to the NAS from other wired LAN systems that are not on the same switch. I’m not sure if that’s important. As long as your internet speed is less than half of your WiFi speed, you shouldn’t really notice a difference.

Medium mode: buy MoCA adapters and use coax. Just be sure to get relatively new ones. They’re generally all 1G minimum, but usually half duplex, so there’s still sacrifice there, but MoCA is generally better than WiFi. The pinch is making sure you stop the MoCA signal from exiting your premise. You don’t want to tap into someone else’s MoCA network, nor have them tap into yours. There are cable filters that will accomplish this, or you can air gap the coax. I’m not sure how much control you have for the ingress/egress of your coax lines. You can yolo it and just hope for the best, but I can’t recommend that.

Hard mode: do ethernet anyways. Usually in rentals, nobody can complain with holes in the walls the size you would get from nails to hand pictures, not much larger than a picture hanging nail, is a cup hook. What I did at my old place, which was a rental, was to buy large cup hooks, and put them every ~18" down the hallway, and load it with ethernet cables. I used adhesive cable runners to go down walls near doors and ran the cables under doors to get from room to room. I got lucky that two adjacent rooms shared a phone jack and I replaced the faceplate with a quad port Keystone faceplate on each side. One Keystone was wired to the phone line to keep existing functionality, the rest were connected to eachother though the wall as ethernet, and I just patched one side to the other (on one side was the core switch for my network). That was my experience, obviously your experience will be different. I used white ethernet to try to blend it in with the ceiling/walls which were off-white. In my situation, I was on DSL and used the phone jack in one of the bedrooms for my internet connection, that bedroom was used as an office and it neighbored my bedroom where I used the jack to jack connections through the wall to feed my TV and other stuff in the bedroom. The ethernet on the cup hooks went from the office to the living room where I put a second access point (first ap was on the office) and TV and other stuff. Inbetween the bedrooms and the living room was the kitchen and the wet wall was basically RF blocking, so I needed an access point on either side, so one in the office near the bedroom and bathroom, and one in the living room, provided plenty of coverage for the ~900sqft apartment we were renting. Most everything was on wired ethernet, and the WiFi was used mainly by laptops and cellphones.

I live by the philosophy of wired when you can, wireless when you have to. Mainly to save WiFi channels and bandwidth for devices that don’t have an easy alternative option like mobile phones and portable computers.

I don’t think you’re in a bad spot OP, and any of these choices should be adequate for your needs, but that will vary depending on what speed internet you have, and how much speed you need for the LAN (to the NAS and between systems).

Good luck.

At home, libreNMS. Just SNMP everything.

For work, whatever the tool of the day is from management.

But if the scammer is using a bot too, then it becomes a null sum, since the bot can have thousands of conversations at a time.

Spam bots should be taken down more than engaged with. If there’s a real scammer on the other end, yes, absolutely, waste that person’s time as much as you can, and as much as you like. People have made entire careers out of trolling them and I endorse it. Scammers are the worst people, taking the hard earned money of his people to try to convince them of a lie just to get their money. This is sometimes true with normal sales, caveat emptor and all that, but when the entire premise of the interaction is based on deception, then to me, it crosses over into scam territory (looking at you, entire duct cleaning industry).

Wasting time making a bot to talk to spam bots is not very helpful. If you can identify that they are not properly filtering their inputs, I would invite you to use an SQL injection and talk to them about little Bobby tables. But by using a bot of your own to talk to spam bots will have such a negligible impact on the harm that scammers have that it’s basically not worth doing. Unless you can scale up your bot to the point of overwhelming the scammers bot into disfunction, it’s not going to provide any real help to those currently being scammed by the bot. Scaling up to the point of getting the bot to malfunction, is also something I would approach with caution, since you have no way of knowing what that limit is, and in the case of cloud systems, the capabilities of the bot may scale far and above what any attack against them could reasonably produce.

If they’re using cloud resources and you can verify that, then there’s a good chance you can hit them financially if you push their bot to its limits since cloud compute resources are not cheap. If you can generate enough traffic for them that the bot scales up significantly, then yeah, you may be successful in forcing the scammer paying for that to shut it down. The trick is doing so without incurring significant costs yourself. It’s still likely, however, that the scammer will simply abandon it (and not pay their bill), and restart the whole thing again later with a new telegram/whatever chat system account later that you won’t be able to track down in a reasonable timeframe.

So it’s somewhat insane to try, it’s easy for them to change the bot to avoid your usage attack and difficult for you to keep track of them and which account they’re using now.

We need to make it globally illegal to run these kinds of remote scam operations, and strongly prosecute anyone doing it. Their ill gotten gains need to be confiscated and sent back to their victims (as much of it as possible), and they should be imprisoned for a very long time.

As far as I’m concerned, this is the way. This is the only way. Legal reprocussions with strong penalties and strong law enforcement of those legalities is the only way to ensure that we crush this trend permanently. Most countries, even those where we see a lot of scamming coming from, have laws that enforce against scams; but the enforcement is very spotty, and IMO, the ramifications of being caught are far too light.

Right now, most civilians don’t really have any good recourse beyond ignoring it. Scambaiters are pretty common and they’re doing good work, even working with law enforcement to get these scammers behind bars, but even that falls far short of the action required to stop such things from continuing to happen. We need strong legislation agreed upon across international boundaries with full task forces to find and prosecute these assholes; we don’t have that, and so it continues.

If you’ve thought of it, they’ve thought of it. Plainly, there are already scam bots floating around, most of the time engaging with them makes it quite clear that they are not actual people, as long as you’re paying attention. Their side oftentimes is completely automated. Get paid send info. The “lifelike” messages they send are canned and only vary slightly from message to message.

I swear, we’ll implement bots to “combat” this stuff and it won’t do anything because it will largely just be bots talking to bots forever. There’s already a nontrivial amount of internet bandwidth consumed by spam email that just gets thrown away as it arrives, now, more and more resources are going to be poured into having bots talk at eachother for centuries without getting anywhere.

I’m a network guy, so everything in my labs use SNMP because it works with everything. Things that don’t support SNMP are usually replaced and yeeted off the nearest bridge.

For that I use librenms. Simple, open source, and I find it easy to use, for the most part. I put it on a different system than what I’m monitoring because if it shares fate with everything else, it’s not going to be very useful or give me any alerts if there’s a full outage of my main homelab cluster.

Of course, access from the internet to it, is forbidden, and any SNMP is filtered by my firewall. Nothing really gets through for it, so I’m unconcerned about it becoming a target. For the rest of my systems security is mostly reliant on a small set of reverse proxies and firewall rules to keep everything secure.

I use a couple of VPN systems to access the servers remotely, all running on odd ports (if they need port forwards at all). I have multiple to provide redundancy to my remote access, so if one VPN isn’t working due to a crash or something, I have others that should get me some measure of access.