JonathanM

It's fine. You may have to cover a couple PCIe pins with kapton tape if your motherboard isn't happy and won't boot, or something like that. You'll know immediately, if it comes up without issue then it's fine as is. Broadcom just wants you to pay for a newer controller instead of bugging them if you have issues with the older one. Pretty much they just put you on notice not to expect support from broadcom AT ALL with that board.

On the UD line item for the disks in question, click on the settings (three gears) and select the passthrough option.

Bad RAM is completely unpredictable, other than the obvious of corrupted data causing random issues. Any data copied to or from a server with bad ram is suspect.

No, with 2 drives in RAID1, the space is limited to just the size of the smallest drive, not 2x. RAID1 means a copy of the data must exist on 2 separate volumes.

You will need to keep Unraid from seeing and trying to access the disks you want to mount. Set up a VM and pass through those disks so you can deal with them there.

Are you sure your vdisk location has enough free space?

Looks like the USB may be corrupt. Put it in another machine and run a checkdisk on the USB stick.

If you shave the top and bottom of the door so there is a 1/2 to 3/4 inch gap when it's closed, convection should take care of any heat buildup. If the drive is actually outside the manufacturers recommended range, then yes, I'd put a heatsink on it. Otherwise, change the global disk alarm temp to match the manufacturer guidelines for the hottest drive, and bump all your spinner drives down to the range you want to see. Ideally you should only get an alarm when something is truly wrong, like a fan gone bad or somebody blocked the closet door.

You don't. Use br0, not virbr0.

Post your docker run command.

No need to rebuild parity, each drive is independent. There are several ways to accomplish this depending on risk level vs time to do the procedure.

I figured as much, but small things like that can torpedo users who don't know any better and blindly follow what's typed. Accuracy is very important.

Unclear what you are getting at. No data drive can be larger than either parity drive, but there are no restrictions on the sizes of the parity drives as long as the data drives are equal or smaller than either parity drive.

Replace one at a time.

Did you find anything like Above 4G Decoding, or Resizing BAR, or Resizable BAR, or Re-Size BAR Support? Also, try toggling UEFI

Possibly a BIOS setting?

nothing in Unraid uses that path

Sounds like a forced shutdown is your only option. The good news is that ReiserFS is very capable of handling that sort of thing, but it will likely take hours for the drives to mount cleanly as ReiserFS replays the transactions. Once you have stable power guaranteed, boot the server and retrieve diagnostics, then start the array and wait for all the drives to finish mounting, which will probably take hours, then collect diagnostics again.

Moving forward with 6.9, you can have multiple cache pools. Using UD for drives intended to stay attached for stuff like appdata is not the preferred method any more.

Poor choice of words. ReiserFS can get super slow sorting out file fragments when asked to write to a drive that's full. That's what I meant by maintenance, it's likely just stuck sorting out where to put things.

I'm more inclined to blame ReiserFS, having experienced almost identical symptoms several years ago. Also, just because those drives have an exponential failure rate compared to similar drives in that time period, doesn't mean that there is a 100% failure rate. In my experience, drives that survive several years are more likely to die relatively normal deaths typically preceded by warning SMART errors. Doesn't mean those drives are good, but without a SMART report you can't pass judgement just yet.

Different ethernet card? Different Unraid version? Try 6.9 rc2

Educated guess here, ReiserFS is doing file system maintenance. Leave it alone for at LEAST 24 hours and see if it resolves. Another issue could be your 3TB seagates, those models are infamous, they have their own wiki entry about the class action. https://en.wikipedia.org/wiki/ST3000DM001

How are you dealing with ventilation?