Hi,

Will one or two of these cards work to make an array out of m.2 nvme ssds? I have a dell r720 with dual CPUs so I have plenty of pcie lanes.

Is there any better solutions for this as I might decide to expand to use more drives in the future.

I have a dell r720 that cannot do bifurcation so is there an alternative?

Thanks

I've been doing some searching and watched some YT vids on the subject, but I want to hear anyone's real world results before I pull components and waste perfectly good thermal paste.

Will replacing two CPUs with one, reducing RAM by half, and replacing the 750W PSU with a 495W noticeably reduce power consumption?

I have a Dell R730 with dual Xeon E5-2650 v4, 128gb RAM, 1x 750W PSU, Quadro P1000 GPU, several drives.

Also, has anyone gotten an X16 GPU to work in a single CPU R730? If not, then I suppose considering going single CPU is moot.

I am sure someone has better Google skills than me, but wondering if anyone has a motherboard that has full lights-out capability, such as BMC for the i7 or i9 processors.

I know this is a DC feature, but want to build a machine around an inexpensive CPU, and XEON and EPYC seem to have (for good reasons) High price points for the same performance as an i9

I have enough space for either a 2u or 3u case, but would prefer a 2u but 2u compute (already have a RS1221+) builds seem more difficult to achieve for my needs. I’d like to be able to run an Intel i9 with a higher end GPU, IPMI motherboard running Windows as its main OS, and also run multiple VMs with a SFP28 interface to my network.

At the end of the day, I’m really looking to learn more about networking, compute and infrastructure in general which is why I don’t want to use something like a Minisforum MS-01.

I’m torn between a Sliger CX2137b (2u) case and CX3152a (3u) case (If another case is better please let me know!)

For 2u, I’ve been looking at this blog post which doesn’t make me very excited about choosing 2u.

At the same time, I feel like choosing a 3u isn't really necessary (but what really is necessary about all of this), but it’s more future proof.

Would love to hear the community’s thoughts on this! If there’s a better place to ask this, please let me know.

I'm planning to get a Digitus wall network rack, probably 9U, or maybe if needed 12U, still deciding.

The idea with 9U is to put small UPS and NAS on the rack's case bottom side itself, but if getting 12U to probably put a shelf for those two, since having more space.

I'm worried about possibly damaging rack's case bottom part with the weight of the UPS, I am assuming it can bend a bit. Should I better get a shelf and sacrifice 1U for it, or will it be fine? I never had rack before so no idea how much weight can case handle.

I would like to access for example im at a race track and i want to add some photos to it is this possible or would i have to be connected to the same network?

My networking can't really handle the 5 tapo cameras I have. Internet and routing comes in via a tplink deco x60 with no useful GUI or CLI. The only benefit of the x60 is mesh wireless - I find it hard to believe that it can support 100s of clients per node. My server sits behind a glinet split tunnel wireguard mini router that is a client of the x60. I use static routes to access the x60 subnet from the glinet subnet. I've used ubiquiti edgerouter, glinet, and lightly played with pfsense as a VM.

But I'm going to be upgrading, the centerpiece being a tower server with 10G and 2.5G Ethernet running promox.

Any advice on topologies or approaches?

I'm liking mikrotik, I need at least 10 Ethernet drops, more if I decide to run some small devices on that rack or run a few more Ethernet runs for 2 port wall plates. (I'm guesstimating about 400 feet of runs).

https://mikrotik.com/product/crs310_8g_2s_in

And would use the sfp+ for the two connections to the main server. All mesh nodes (a mix of x60, m4r, whatever) would be on access point mode, most via Ethernet backuaul, which I think still gives me mesh networking - there's going to be five of these things. And the mikrotik would be running wireguard to access my lan. Then I think another

I guess the alternative would be a udm and ubiquiti ap for a single ecosystem. The x60 is end of life and I grow tired having routers behind routers. But like I said, I like the hacky flexibility of mikrotik.

Hi,

so after reading for few months, im about to make the jump and go down the rabbit hole. Starting easy and maybe slowly, later i would love to add a NAS.

My question is mainly about the device that it will run on. Is this HP Elite Desk 800 G6 good starting point? Was thinking about minis forum or beelink, but after seeing random deaths and some sort of problems with them, im not sure anymore.

Price: 360€

CPU: Intel Core i5 10500 - 3,1 GHz (6 core / 12 thread)

RAM: 16 GB

SSD: 256 GB

I've been building my own PCs for about 20 years now, and just last week, I encountered a problem I never encountered before, and thought I'd share my experience.

I bought a used mobo/CPU/RAM combo from eBay some months ago to build a home server, only now got around to testing it and setting it up. Supermicro X9SRL-F, Xeon E5-2690 v2, 128GB Samsung ECC RAM. Nice stuff. Step one was slapping it on a test bench, hooking up a power supply, keyboard, monitor, and running memtest. Everything was great, no issues. So I moved on to installing everything inside a case (specifically a Phanteks Enthoo Pro 2, great case), additional add-on cards and etc, and eventually it was time to power it on. Buuuuut it wouldn't boot. Took out all of the addon cards I hadn't tested yet and tried again, still wouldn't boot. BIOS was giving me some error codes that, upon Googling, seemed to suggest a problem with memory detection.

Weird, I thought, considering it just the day prior fully passed several memtest rounds. Did a little more digging and saw some advice suggesting that a lot of people fixed this error by reseating all the memory as well as the CPU. I thought, fair enough, this is 10-year-old server stuff, probably good to do that for a variety of reasons. So I took off the cooler, cleaned it all up, removed the CPU, cleaned it top and bottom, inspected the motherboard for any bent pins or stray thermal paste. No bent pins, but I did see a small piece of some unknown debris in there among the CPU pins. Don't know what it was or if it was in fact the culprit, but whatever it was, I removed it. Reseated the CPU, new paste, mounted the cooler. And during all this, I also removed all the RAM sticks and reinstalled them in reverse order so that every stick was in a different slot than before. Tried booting up again aaaaaaaaaaaaaand the memory error codes still persisted.

I was still confused as to why it passed memtest just fine 24 hours earlier but the motherboard wouldn't even let me boot up memtest anymore. Started removing RAM until a sufficient amount was removed to cease the error codes, which in this case were the sticks populating the two RAM slots nearest the top of the case. I then memtested just those two sticks of RAM that were causing issues in different slots, but they tested fine. So I concluded, okay, maybe it's just those two RAM slots are dead. This is a used eBay motherboard after all, maybe this is why they were selling it and didn't disclose the issue.

But I was still bothered by the idea that it all memtested fine before installing it in the case but the top two RAM slots were dead after installing it in the case. And then after some more Googling, I found someone from six years ago on the TrueNAS forums with my same model motherboard with my same issues, and they eventually discovered and fixed the problem.

What was the problem?

The case had pre-installed standoffs for motherboard installation, and it turns out that one of the standoffs that was installed but not used by this particular motherboard was in juuuuuuust the right place to make contact with and short out some of the RAM slot soldering points on the back of the motherboard and cause electrical issues. So I removed the motherboard, removed that one particular standoff and all of the other preinstalled and unneeded ones just in case, reinstalled all my hardware, booted up, and whaddya know, no error codes anymore, ran memtest with all the sticks again and it all passed just fine, the machine was back to working like it should have been all along. All of that head-scratching and puzzlement and thinking I had faulty hardware and got shafted on eBay, when really it was just a unique variety of user error.

It's nice that case manufacturers will sometimes preinstall some commonly used motherboard standoffs for general users' convenience, but in this case, it turned out to be quite inconvenient for me! It was very easy to fix once I discovered it was these causing the issues, but I was very close to assuming I just had a faulty motherboard or RAM when in fact everything was perfectly functional.

So yeah! If your PC case has any preinstalled motherboard standoffs, it turns out it's good practice to remove any unneeded ones. Never had this problem before, but now that I've had it once, you can be sure this is something I'll do with every build in the future. It's funny, though, because it makes me think of how many people must be RMA'ing new hardware that appears faulty, when it turns out it's perfectly fine hardware that was acting faulty because of user-related reasons like this. Similarly, I've had so many new PCs not boot the first time because I overtightened the screws on the CPU cooler and the motherboard was being flexed in a bad way. Backed the CPU cooler screws off a half-turn or two and then they all booted fine in all those cases for me, but someone else may have just assumed it was a DOA CPU or motherboard when in fact it was user error.

Food for thought. But at the very least, I hope this tale prevents someone else from wasting hours of troubleshooting in the future.

Hi Labbers,

Longtime lurker, first time poster... I purchased a DL380 G9 which came with the rails but only the side of the rails that attach to the actual server. From my research, it seems like this part of the kit is the most important part as it has the holes in the right place for attaching to the server... my question is... Does anyone know if this will just slide into a generic rail kit if I just take the outer rails and discard the generic inner rails? And does anyone have an recommendations for some generic rails that they think would fit well with OEM inner rails?

Long shot, I know but any advice would be much appreciated. Unfortunately my rack is in a spot that gets a bit of dust so i need to be able to slide the server out and give it a blow out every 6 months.

I have a plan for two systems ( + a synology NAS for taking backups/NFS in the future) that I want to host some basic video game servers on (terraria, minecraft, etc), all of which I want to be public facing. Yes, I'm aware routing them through a VPN is more secure, for my use case I will not be using one.

My network/security plan so far:
Cloudflare reverse proxy --> Modem --> UXG-Max (running IDS/IPS, the Unifi controller is running on a Pi on the main LAN) --> segmented VLAN (only has access to the Cloudflare rp IPs) --> 1st system* (docker with a local Traefik reverse proxy and Authelia for 2fa into proxmox console) --> Proxmox system (connects to the Traefik rp and runs the game servers in Ubuntu VMs)

Questions:
1.1. Is having the auth server, certs, and reverse proxy on the "1st system" best practice if I wanted to upgrade the single Proxmox system into to a cluster in the future?

1.2. If so, would an 8gb Ras pi 5 be fine for what the "1st system" will be running? Or will I need a 'beefer' system?

1. How should I monitor my "1st system" and "Proxmox system" for attacks? A SIEM like Wazuh?

2. Whats the "risk" of this setup? (I know, kinda stupid question. I'm looking for a "this is so dumb, don't ever do this" or a "if you actively monitor, patch, and have backups ready to go you'll be fine" sorta answer)

3. Anything I'm missing?

As the title implies, I'm just looking for advice here, any is greatly appreciated. I would love to hear if you're running a similar setup successfully!

Thank you for reading

I've been trying to set up my home lab and I need help configuring my NIC. Here is the one I bought: https://www.amazon.com/dp/B074PQWQB5?psc=1&ref=ppx_yo2ov_dt_b_product_details . I'm trying to separate network traffic on my Intel Gigabit ET Dual Port Server Adapter using Windows 11 Pro and VMware in my home lab. I need one port dedicated to my main PC for gaming and browsing, and the other port for my hacking lab VMs running on VMware. Despite having a Cisco managed switch with VLANs configured, Windows 11 doesn't support Intel PROSet and ANS for VLAN tagging, making it challenging to achieve this separation. I'm looking for a way to ensure each port handles traffic exclusively for either Windows 11 or VMware to maintain network isolation and performance without using costly third-party tools. Any advice or solutions would be greatly appreciated.

So last year I have run ubuntu server on my HP elitedesk. It was running:

• portainer

• Gitea

• Plex

• Technitium

and it was working really nicely. I was also running some devops tests before implementing it in my job.

But recently I have found need for having running windows only apps. Running these apps on my M2 macbook is quite boring and I could benefit of being able to connect from everywhere.

So I have installed windows server 2022 and run Microsoft Remote Desktop. It all works so well. But then I have check the price tag and I was like whaaat.

Now I think that the windows server is quite not needed and I am thinking of running just windows 10 pro on the server.

What would be your approach?