Here's a quick and easy step by step guide to getting Simplemining OS to run properly with 12 GPUs on the Biostar TB250-BTC Pro motherboard if you're encountering PCIe Bus Error messages.
The problem:
You've hooked up your cards, you've checked your risers, everything seems to be in order, but as soon as you boot up you're flooded with scrolling PCIe Bus Error messages that prevent your rig from even booting up (or cause it to boot extremely slowly and inhibit mining after booting). They probably look something like this -
PCIe Bus Error: Severity=Corrected, type=Physical Layer, id=00db(Receiver ID)
device [8086:aZee] error status/mask=00000001/00002000
[0] receiver error (first)
If this is what you're seeing, fret not! While I'm far from an expert on linux kernels and PCIe bus errors, what I've come to understand about this issue is that these error messages belong to something called Advance Error Reporting which essentially warns you when a PCIe port is behaving strangely and MIGHT fail, but hasn't yet. Whatever error was thrown was quickly resolved, but the system wants to warn you about it. The problem with these messages is that they cause a micro-pause during the system boot, and when several of them are thrown each second during the boot process, they effectively paralyze the boot. Obviously that's not cool when we want to be up and mining ASAP.
After communicating with the smOS dev, who was super helpful during the whole troubleshooting process, and through a few days of research and testing on our end, we eventually found the silver bullet to put an end to this issue and get our 12 card rig up and running with no issues to speak of.
And without further adieu...
The Solution:
In our experience, this error was only thrown after connecting more than 2 GPUs, so first hook up 1 or 2 GPUs and boot into Simplemining OS. If you're like us, you won't get any errors, and once the OS is finished updating etc, you should begin mining normally. Now you need to get into a terminal and make a small change to your Grub config file.
- Press CTRL + ALT + F3 to open up a new terminal window
- Log in as root (I won't post the root password here, you can either get it by emailing admin@simplemining.net or with a little Google-Fu)
- Enter "nano etc/default/grub" (without the quotations)
- Change GRUB_CMDLINE_LINUX_DEFAULT="quiet" to GRUB_CMDLINE_LINUX_DEFAULT="quiet pci=noaer" (this turns off Advance Error Reporting)
- Press CTRL + X to save
- Select Yes
- Hit Enter to confirm the file name
- Enter "update-grub2"
- Enter "update-initramfs -u"
- Shut down the rig
- Add the rest of your cards
- Switch the rig back on
The rig should boot right up normally and begin whirring away with all 12 cards. Credit for this solution goes in large part to the Simplemining OS admin as it was his initial suggestion and instructions that got us going on the right path. Huge thanks to him for the prompt and personal support. If you have any other issues or this solution doesn't work for you, feel free to leave a comment and we'll see if we can figure out the issue together.