archomrade [he/him]

  • 3 Posts
  • 19 Comments
Joined 1 year ago
cake
Cake day: June 20th, 2023

help-circle





  • There seems to be two types of homelabbers with regards to storage:

    • Those who take storage redundancy seriously
    • Those who don’t seem to care

    I’ve made the mistake of asking the second group what they thought about types and quantities of storage, and I got quite a few “why are you concerned?” type questions. My guess is that they regard obtaining data to be free/trivial, so storing it redundantly is a pointless cost. I’ll just say that I don’t share their cavalier attitude.

    This setup is my personal goal, and I think refurbished drives are the best way to go about it (provided they are reasonably taken care of). If you’re working in a redundant setup, the age of the drives matters a lot less.


  • The only real concern I could see with these is a nefarious actor having access to your real-time usage to determine your habits. I guess the argument can be made that the POCO is spying on you, but I really don’t see a benefit to them knowing your usage aside from network/grid management. Perhaps a locality can use this data to determine if you’re growing weed, but other than that, in a world where all of our information is harvested anyway, it seems pretty inconsequential to me. if you’re that concerned, get a solar/battery system and turn off your main breaker.

    That, and be able to adjust billing to surges.

    The secondary concern is reading the real-time usage by a nefarious actor, which has proven to be useful to ML applications for intuiting what types of activities are happening on the circuit. Obviously this pales in comparison to the potential abuse allowed by smart NICs and modems, but as a secondary measure or where a user is obscuring their network activity with encryption or similar, a real-time power reading would be helpful too.

    It’s just another vector that enables possible abuse, and one that you don’t have any choice over. My utility installed one of these on my house this year and we were not given any choice.


  • If I could trust an authority to create effective regulation, and trust the company to follow it, then I wouldn’t mind the technology itself.

    I personally don’t buy that personal meters provide much more useful data about grid performance than if the continuous meter was placed further up the line (maybe at junctions and transformers?). To me, the risks far outweigh the benefits. The push for these individualized meters is really unsurprising though, since it provides far more financial benefit to them if they can measure each user’s energy use and tailor pricing to that individual. Even if they could guarantee that data never left the utility, its existence is still problematic just by the nature of the oppositional relationship between utility and the user.




  • I think this might strike the right balance.

    I kinda get where everyone else is coming from, about enterprise grade drives being more performance than needed, but with the reliability of these drives it just doesn’t make sense to me to get rid of them all. The power use difference between these and other NAS drives is pretty negligible, and the parts needed to utilize a SAS array aren’t all that expensive in comparison to buying new drives (even if I sold all of these, I don’t think i could get as much for them as what even a used consumer drive would be).

    These are 4TB drives, so an array of 4 in a raid 5 configuration would get me 12 TB. I could go up to a more redundant raid with more disks, so maybe I’ll get a shelf/case that can hold 6 or 8 or even 12, and plan for adding more later. I have a high-efficiency micro-atx sitting on my desk with a couple empty PCIE slots, all i’m missing I think is a SAS controller and a case/PSU.

    I’m always hurting for more space, so to me it’s smart even just to have them sitting on a shelf for when I need another.



  • This is a great observation, and it made me do some math:

    If my point of comparison is something like a seagate ironwolf 4T vs a WD Ultrastar 4T:

    Seagate Ironwolf: 
    - 3.7W*24 Hours/day*365 days/year = 32kWh per year * $0.18/kWh = $5.84 per year in power usage * 12 disks in an array = $70.02 per year
    
    *Edit: Looking at this closer, a more reasonable comparison would be an ironwolf PRO disk, since this is a NAS use-case (24-7 run time, large and repeated writes and reads, ect). The power consumption for that is 5.5W, which is a lot closer to the Ultrastar*
    
    WD Ultrastar:
    - 7W*24 Hours/day*365 days/year = 61kWh per year * $0.18/kWh = $11.05 per year in power usage * 12 disks in an array = $132.6 per year
    

    Seems like i’d save maybe $70 per year. I feel like that difference might even be justifiable if the enterprise drives are half as likely to fail (seagate ironwolf has an AFR of 0.87%, WD Ultrastar is 0.44%).

    Something to think about, at least








  • I second this. It took me a really long time how to properly mount network storage on proxmox VM’s/LXC’s, so just be prepared and determine the configuration ahead of time. Unprivilaged LXC’s have differen’t root user mappings, and you can’t mount an SMB directly into a container (someone correct me if i’m wrong here), so if you go that route you will need to fuss a bit with user maps.

    I personally have a VM running with docker for the arr suite and a separate LXC’s for my sambashare and streaming services. It’s easy to coordinate mount points with the compose.yml files, but still tricky getting the network storage mounted for read/write within the docker containers and LXC’s.