Viceroy's recent activity
-
Comment on Looking for feedback on a homelab design in ~tech
-
Comment on Looking for feedback on a homelab design in ~tech
Viceroy Link ParentThis is such an excellent write up, thank you for taking the time to type it all out! I absolutely will need some help with the zfs setup so when I get to that stage I will certainly be reaching...This is such an excellent write up, thank you for taking the time to type it all out! I absolutely will need some help with the zfs setup so when I get to that stage I will certainly be reaching out.
Roughly how much data are you expecting to store?
I know that a 90-bay or 45-bay storage is a lot and I am the first to say that I will likely never use the entire capacity, but I have been burning through ~40TBs embarrassingly quickly and I thought it would just be a good exercise to think through what a true “future proof” system would look like. Also, I usually see prices for JBOD cases to be pretty much negligible between larger drawers, probably since most of the cost of a system like this is the drives themselves, it never struck me as practical to limit purchases to smaller bays.
I absolutely will be buying recertified or refurbished drives. I don’t see any reason to shop new.
The vdev configuration is less about increasing for more throughput and more about chopping up the drives in a way that makes sense with quite a bit of online literature essentially pointing away from making extremely wide vdevs. I had thought about doing 6-wide z1 vdevs but that alternative seems more risky even though the overall parity is roughly the same. The only real downside I see for my setup is that I need to buy 11 drives for each new vdev which is quite a lot. I’m no expert here though, happy to hear your thoughts, I was using this tool to help visualize failure rates: https://jro.io/r2c2/
On RAM, that was exactly my thought. 64GB DIMMs are way too expensive for whatever reason so I was thinking of starting with two 32GBs and then expanding as necessary.
Thanks for the information on the Supermicro motherboards, I have definitely been thinking Rome would be a little too out of date, but Im glad to hear others are thinking the same.
SLOG and L2ARC: This information is invaluable, thank you for the write up. Finding objective and nuanced information online about SLOG and L2ARC has been… challenging to say the least. It seems to be polarizing in a way that doesn’t make sense.
- For L2ARC, this was my gut instinct. Relatively minimal cost for decent gains overall. Thanks for confirming
- For SLOG, this is great information. I will probably keep things running synchronously with a SLOG drive. I was completely unaware of the PLP functionality, and will definitely be something I will look out for when it comes to picking parts.
-
Comment on Looking for feedback on a homelab design in ~tech
Viceroy Link ParentI really appreciate it! I definitely will be reaching out as a get more into the Proxmox setup, but since I am still in the early stages of deciding how to set everything up I am still a bit away...I really appreciate it! I definitely will be reaching out as a get more into the Proxmox setup, but since I am still in the early stages of deciding how to set everything up I am still a bit away from buying anything. I would definitely be down to set up some mutual off-site backups when I get around to that point, within reason of course.
-
Looking for feedback on a homelab design
I wanted some help with a homelab server I am in the beginning stages of designing. I am looking for a flexible and scalable media and cloud system for home use, and I thought this community would...
I wanted some help with a homelab server I am in the beginning stages of designing. I am looking for a flexible and scalable media and cloud system for home use, and I thought this community would be a good place to source feedback and recommendations before taking any real next steps! I really want to check that I am approaching the architecture correctly and not making any bad assumptions. I am open to all feedback, so please let me know what you think!
I already run a simple home server and I have typical homelab FOSS apps, such as jellyfin, navidrome and audiobookshelf, but I am also interested in migrating away from cloud storage using nextcloud, immich, etc. In an ideal world, this setup would also allow me to leave windows on my main machine and use a windows vm for business related work that can’t be done on Linux. I will likely be the one primarily using the services, however I could expect up to 10 - 20 users eventually.High level setup is with two machines:
- Proxmox Server
- TrueNAS Scale server
- JBOD with either 90 bay or 45 bay storage
- 10G switch
This might be a stupid setup right off the bat, which is why I wanted to discuss it with you all! I have read a ton about using TrueNAS as a WM within Proxmox, but I just like the idea of different machines handling different tasks. The idea here would be to set up the TrueNAS server so it can be optimized for managing the storage pool to allow for easy growth. While the Proxmox server can handle all the VMs and connecting users, with higher IO, etc.
TrueNAS System Specs:
- AMD ryzen CPU and motherboard
- 64 or 128GB ram
- Mirror 500GB M.2 NVMe OS Drives
- GPU if necessary, but hopefully not needed
- Dual 10gb pcie card if the motherboard doesnt already come with them
- An hba for the JBOD something like the LSI SAS 9305-16e
- SLOG and L2ARC as necessary?
JBOD enclosure
- While I am interested in a 90-bay enclosure, I would only realistically be starting with two vdevs which is why I think a 45 bay enclosure wouldn’t be an issue.
- Im tentatively planning for an 11 wide Raidz2 vdev configuration. This would hopefully scale to 8 vdevs with 2 hot spares or 4 vdevs with 1 hot spare.
- All drives would be HDDs
Proxmox Server Specs:
I am less familiar with the specs I will need for a good Proxmox server, but here is what I am thinking.
- AMD epyc and motherboard if I can get my hands on a less expensive one. Otherwise I was thinking a higher end AMD ryzen cpu
- 128 or 256GB ram
- Mirror 500GB M.2 NVMe OS Drives
- Somewhere between 2 and 8 TBs of SSD storage. Depending on the number of drives, I think this would be a single drive, mirror or raidz1.
- This storage will be used for all the vm configuration and storage, except for something like Nextcloud where the main storage will go onto the TrueNAS mount.
- I would also use this for temporal storage such as downloading a file before transferring it to the TrueNAS mount.
- A dedicated GPU primarily for transcoding media streams, but also for testing and experimenting with different AI models.
- Dual 10gb pcie card
Questions:
- I know Proxmox can do zfs right out of the box so I know I don’t need the TrueNAS server, but splitting it this way just seems more flexible. Is this a realistic setup or would it just be better to let Proxmox do everything?
- Does anyone have experience creating NFS shares in TrueNAS for mounting in Proxmox? I would be interested in thoughts on performance, and stability among any other insights.
- Do any of the system specs I listed seem out of line? Where and how do you think things should be scaled up or down?
- If I ever did expand to a second JBOD shelf, assuming the first one was full first, is it be possible to create new vdevs that spanned across the shelfs without losing data?
- Is SLOG and/or L2ARC necessary for this setup? What capacity and configuration would be best?
- What else have I missed?
Lastly, a quick blurb:
I have been building PCs for a while and undertook building a home server a few years ago. I loved the experience of learning Linux (the server is running Ubuntu), picking up docker, and learning more about the FOSS community has been a joy! Part of this project is to learn along the way but also have a setup that I can build towards over time! Proxmox, TrueNAS and zfs would all be new to me so I really see it as an opportunity to explore. I want a solid media and cloud server setup, while also giving myself the freedom to explore new operating systems and general hypervisor functionality.
22 votes -
Comment on Tesla Cybertruck owners shocked that tires are barely lasting 6,000 miles in ~transport
Viceroy LinkAs others have said im sure the light truck tires on a heavy vehicle with instant torque from the electric engine is probably a majority of the issue, but I wonder if the Cybertrucks ability to...As others have said im sure the light truck tires on a heavy vehicle with instant torque from the electric engine is probably a majority of the issue, but I wonder if the Cybertrucks ability to change toe angle contributes to the faster wear by grinding the tires at low speeds. I could see a lot city driving around parking lots and such increasing the wear at low milages.
-
Comment on <deleted topic> in ~finance
Viceroy Link Parent12-Month CDs reflect the current rate environment, which is relatively high. Longer term CDs aren't just financial institutions 'betting' rates will go down, it is statically likely that they will...12-Month CDs reflect the current rate environment, which is relatively high. Longer term CDs aren't just financial institutions 'betting' rates will go down, it is statically likely that they will return back to the average, an effect known as "reversion to the mean". So, long-term rates will always price in this anticipated pull toward the average, making them less sensitive to the feds rate changes.
-
Comment on What AI tools are you actually using? in ~tech
Viceroy Link ParentThis is a perfect example of what generative AI is great for. I work in the AI space, and I have always hated the clickbate headlines about it. AI has some amazing possibilities and potential, but...This is a perfect example of what generative AI is great for. I work in the AI space, and I have always hated the clickbate headlines about it. AI has some amazing possibilities and potential, but the current applications are much more mundane. It really is best used as a jumping off point for more research, summarizing text/code to digest easier or basically doing the heavy lifting in brainstorming so that a person can then expand on the generation with their experience.
-
Comment on Advice on expanding storage in starter homelab/media server in ~comp
Viceroy LinkI decided to build a home-lab/media server a couple of years ago as a project during the pandemic and I decided I wanted to build the entire thing from scratch so I went in a slightly different...I decided to build a home-lab/media server a couple of years ago as a project during the pandemic and I decided I wanted to build the entire thing from scratch so I went in a slightly different direction than most of the commenters here. I'll try to provide a brief overview and if you have any additional question just let me know!
Intro:
First, this was mostly a learning project for me, I don't have a formal programming or hardware background so it was primarily to challenge myself to learn newer technologies. I was specifically interested in learning linux, docker and networking configurations so instead of using something like UnRaid I just used Ubuntu. I specifically targeted uses cases of streaming media and document storage so I had a pretty good idea what I needed from the hardware. The goal was for the server to always be on so I didn't bother including any type of wake scripting, since it wouldn't be needed.Hardware:
Roughly here was the build: Micro-atx case and mother board, 16 GiB of ram, 10th gen i5 intel cpu, one Nvme for the OS and four 4TB Ironwolf drives for storage. At the time this cost around $900 for everything.Software:
I loaded the most recent version of Ubuntu and off I went. I decided that I didn't want to go headless since I still wasn't that comfortable with linux so I went with a standard install. Then I implemented a software raid on the storage drives, I went with raid 5 as it seemed to give the best tradeoff between redundancy and efficient drive use. Since docker was one of the things I was keen on learning all of my services are dockerized, and I manage all of them through Portainer. I use Jellyfin for movies/tv, Navidrome/Symfonium for music and nextcloud for document storage. I was also interested in being able to access my services externally so I also bought a cheap domain and setup a reverse proxy for the external connections. Lastly, I use wireguard to securely access the services that are not exposed externally. I also put in several more containers for vpns and sourcing content, but if you have questions about that then just DM me.Thoughts:
Overall it was a great project and I have a lot of fun with it. On the 'expand-ability' front I really like this setup, docker makes it very easy to spin up/down new services I want to try without interfering with my main storage or UI applications and the storage can be easily expanded.Happy to answer any questions you might have!
-
Comment on Is anyone here a consultant? I have questions... in ~life
Viceroy LinkHey, something I can try to answer. I have been a consultant independently, as a part of a small firm and a large firm. At a very high level, consultants are hired to solve problems or implement...Hey, something I can try to answer.
- I have been a consultant independently, as a part of a small firm and a large firm.
- At a very high level, consultants are hired to solve problems or implement solutions that a company does not have the internal resources to address. They are usually hired to deliver a specific outcome or meet some delivery in a set amount of time. You should expect to work hard to meet this goal but you should be compensated well for achieving it. Typically, people will treat you as the expert on the subject at hand, so you should know your stuff if you are going to take the job.
- This depends on firm size, etc. But for larger firms, billing rates are typically 3 - 4 times the salary of the employee. As an independent consultant, I would recommend billing at an hourly rate at least twice what you expect to make in a year. So, for example, if you wanted to make $100k a year, which is roughly $50/hr, I would recommend charging at least $100/hr. Rates can be as high as you think someone is willing to pay you for your work so if your expertise is extremely niche your hourly rates can be higher to compensate for the difficulty a firm would have trying to hire that skillset.
Happy to answer any other question you might have.
-
NordVPN changes to username and password encryption cause Auth_FAIL in OpenVPN/Gluetun
Recently NordVPN rolled out an update which forced users to use an encrypted username and password combination when connecting through OpenVPN. I haven't seen any posts on this here, and it took...
Recently NordVPN rolled out an update which forced users to use an encrypted username and password combination when connecting through OpenVPN. I haven't seen any posts on this here, and it took me way longer than I want to admit troubleshooting this issue because I knew my original credentials were correct.
If you use a gluetun container for routing any of other containers traffic, you might have recently noticed a 500 Internal Service Error in your Health Status and when you check your logs you will find a AUTH_FAILED message.
Solution below:
- Go to NordVPN website and log in (using your normal credentials)
- Under accounts, services, click NordVPN
- Click "Set up NordVPN Manually" at the bottom of the page
- You will receive an email verification code, using whatever email you have set up for your NordVPN services. Type this code into the popup window.
- Copy your new encrypted credentials for your Open VPN client settings.
This is my first post, please add tags as required.
24 votes
I definitely respect this opinion, I have had my fair share of "driver issue causes complete system meltdown leading to multiple days of troubleshooting", but I will say this is mostly for fun, and at least for now, the data and services I'll be running aren't mission critical. For example, I completely rejected the idea of hosting Vault Warden on my current server because I didn't want my entire password manager to go up in flames if I didn't configure something correctly.