whipdancer

Docker gives your container a randomish name if you don't name it. If you get into them, you can get some idea what they are. You can try checking the docker logs.

Check if anyone has had issues with the MB chipset you're looking at. Same with CPU and Memory. I just repurposed my 10 year old Dell laptop (Latitude 7390, 8gig ram, i5-8350U cpu, busted screen) to be my Frigate NVR. I'm running 6 cameras (2x 1080, 4x 4K) with a Coral TPU and the Intel UHD620 iGPU. I mapped shares from my Unraid server to the laptop (using NFS) for storage needs. No issues, detections working great, notifications being sent. I had thought about buying something, but thought I'd give this a try. Glad I did, and doubly so, because the iGPU in my server can stay dedicated to Plex.

I use Unraid for my home and as a NAS. My homelab is separate. My homelab uses Unraid for NAS storage, but runs its own stack of containers. I expose a very limited slice of Unraid using tailscale now. I used vpn + reverse proxy previously. Frankly, if you're to the point where you want to use Unraid in a way that they expressly warn not to do ("don't open Unraid to the internet") and you're worried about vulnerabilities - then you need to use Compose with containers that meet your standards. You can start rolling your own containers so you can lock them down. Getting it right and keeping the app you're trying to deploy working correctly isn't always straight-forward or easy. I write some docker files on my own, that I keep in Github. I've never tried to automate maintenance in Unraid (other than the plugins that are available). I use scripts and chron jobs on my homelab server that help to manage containers running there. I use dependabot on my github repo where I keep all my docker files. If I can find something like Snyk to also scan my repo for free, I will add that. I'm looking at adding a Claude review bot to my repo, which would also provide CVE checking. Claude/ChatGPT/whatever can give you pretty solid guidance on how to write a docker file. If you don't use those tools often, then start with OpenAI Academy and one of their beginner tutorials. Learning the basics on how to use them is the difference between solid, basic guidance and hallucinations of unicorns doing it for you. Something you need to accept & remember - you can't control vulnerabilities in the App you use, regardless of what container you use. If something in the *arr stack or immich or whatever has a vulnerability because they use a library or package that is vulnerable - you either quit using that app or you upgrade once they've addressed it. The truth of using nuget/npm/pypi/gems (or any other public source of code) is that there is a chaotic mess of upstream entanglements for a lot of apps we use. You can't fix those issues. You can only choose how you respond to those issues.

What's your question?

I've only encountered this once. It was a cheap, no name SSD I bought on Alibaba. It was recognized by my server, but the information didn't match anything on the label. The information my server reported was correct.

Is there a specific issue you are trying to address?

You’re mistaken here. Commercial just offers you someone to will ignore you or miscommunicate for weeks on end and still get no actual help. Software companies like Unraid manage to split the difference, with actual help coming from both sides, but you still will have to handle the work yourself.

It is highly unlikely that using a molex to sata adapter killed anything. Something else is wrong. If you have another way to test each of the drives, that will tell you for certain if they are dead or not. I keep a usb dock connected to my regular pc that allows me to quickly determine if a drive is (most likely) dead or not.

How do you know your 3.3v mod is still good? Does the system you tested with use the same power supply as your new setup? Do you have a molex to sata power adapter you can test with in your new system?

I shop on price, so I have a mix of Seagate, Western Digital, HGST, Toshiba - but mostly Seagate.

As it turns out, both of the drives were on a single power splitter. Reworked the power cables a bit and running sync to rebuild everything. I'll update once complete.

I had 1 drive in the array suddenly show up as "device is disabled". I pulled diagnostics, rebooted, ran a short Smart-Test, no issues show up. Dropped it out of the array, then added it back and started a sync. Then I had a different drive in the array suddenly show up as "device is disabled". I canceled the sync, pulled diagnostics, and kicked off a long smart-test on both drives. I'm attaching both sets of diagnostics here. This is a new build that I just completed. New PS, cables, case, etc. The first drive was in use in my prior build. The 2nd drive is a warranty replacement that just came from WD. Both were precleared (x3) since neither had data to worry about on them prior to going into the array. I've looked at the logs but I don't think I know enough to be confident about what's going on. First_bluesmaster-diagnostics-20251020-0944.zip Second_bluesmaster-diagnostics-20251020-1052.zip

They sound like they are treating it as if they are building a regular desktop PC. That's the wrong approach. They should test the components by using a trial of Unraid, booting from a thumb drive as you would.

Any recommendations for quality SFF cables for an HBA? I've purchased from Amazon previously when I was running 2x Dell 310 (8x internal) hba's and never had an issue with cables. I'm now running a 9300-16i (all my hba's have been purchase from Art of Server). The mini-SAS to SATA cables that I have seem to all be of similar quality, but I've recently gone thru 2 replacements trying to clear up UDMA-CRC errors. I finally got the errors to stop, but it made me wonder if anyone has any feedback on what makes a good cable or advice on how to treat them/route them to avoid causing damage.

Side panels off can hinder airflow if its a quality case (the better cases are actually designed with airflow in mind). I printed a fan mount and added a fan to my HBA just so I wouldn't ever have to worry about it.

From what I've been learning about Proxmox, core count seems to be very important. I wouldn't count on the G8505 being sufficient to run everything you've mentioned virtually, but that's just IMO.

I have no idea if they are meant to be used sparingly or not - I've never heard that before. A stress test is used to help identify any immediate issues. The assumption is that if there is a weak link present, 300+ hours of constant use could/should find that weak link and cause it break, hopefully. Completely anecdotal at this point as I don't know of any large study done on the matter. Backblaze does a basic prep of a drive before it goes into service and they are the only group I'm aware of that both tracks and publishes their stats. I use any of a number of basic tools to do a surface test. Same with the DOD wipe. A google search can point you to a bunch of options. It mostly depends on what you have available to do the testing. I use an old windows desktop to do the surface test and DOD wipe because the wipe seems to tax system resources enough that it interferes with normal use. Plus it takes about a week to do, so I need something that I can leave alone for a while.

I rarely buy new drives unless they are a great deal. I primarily focus on used enterprise drives (just added 4x 26tb Sea Exos drives). They come with a 3 year warranty handled by AllState. I just had to RMA 3x 18tb WD Red Pro (bought new 3 years ago) and 3x Sea Exos 18tb drives (bought used 3 years ago). The WD are all being replaced by WD. The Exos were all refunded. I’m still waiting on WD as only 1 replacement drive has arrived so far (about 3 week turn-around at this time). The quick refund on the Exos is why I already have replaced them with the 26tb drives. I no longer do advanced replacement with WD, because they lost one of my RMAs a few years ago and it took over 6 months to get the full retail price charged to my card reversed because of their mistake.

I couldn’t find any scandal about Seagate selling used drives as new, but I did find a bunch of complaints about 3rd parties selling used drives as new. That’s got nothing to do with Seagate and everything to do with knowing who you are buying from. According to Forbes and USA Today articles - It’s foolish to think that only buying WD Gold or Red Pro guarantees anything. I bought 4 Red Pro. 3 were exchanged under warranty in the first 18 months. That’s why my bottom line is $/TB and mostly buy used enterprise drives.

I use a mix of enterprise drives, nas drives, consumer desktop drives. All have worked without issue. Examples from all categories have failed prematurely. There are specific types of drivers that are less performant or less optimal (SMR), but even those will still work. I have removed all SMR drivers from my system, but that was because I wanted to maintain a specific level of performance, not because it had to be done. Outside of your specific aesthetic requirements and tolerance for warranty requirements, I don’t think it actually matters that much. I shop based on my minimum size requirement and $/TB, which leads me to frequently buy used enterprise drives.

I use a Firewalla device to do this.

They should be fine. No. Unless I'm mistaken, they are manufactured by WD and sold with HP's private label on them. Some people swear by WD, others by Seagate or Toshiba. I buy which ever is the best deal based on my needs and my target HD size. Right now, any drive I buy needs to be at least 20TB for it to make sense for my setup (I'm after density, not drive count) - so whoever has the best deal on a drive of that size is where my money goes. I probably buy used drives from ServerPartDeals more than anything else - because they tend to be the best value. My anecedote on drive reliability - 4x new WD Red Pro - total use now going on 4 years, but 3 were replaced under warranty within the first 12 months 3x new Seagate Exos- coming up on 2 years no issues 4x used Dell (WD) HC drives - 1 replaced in the first 3 months - going on 4 years since with no issues 2x used Seagate Exos - going on 1.5 years no issues I'm currently stress testing 2x Seagate external 26TB drives that I plan to put into my system. Before that happens, they will have had a full surface test + DOD Mil spec 3 pass wipe done. If the smart reports are clean once all that is completed, I will shuck them and 1 will go into my system, the 2nd will become a ready spare drive.

My guess would be it's a fake based on the seller and price. I doubt the seller is knowledgeable enough to know how to determine authenticity, nor interested in doing so (assuming the seller has ever actually handled the merchandise) since they obviously do not special in computer related equipment. I also don't think equivocating over $100-150 when its something that should have a lifetime use of 7 to 10 years is worthwhile. All that said, doesn't mean it won't work.