Topic: What a nightware ... Lesson learned. - page 2. (Read 544 times)

That miner life! I feel your pain man. But guess what, bet you'll never have that problem again!

I remember my old IT guy would always ask my guys when they had computer issues "Is the cord plugged in all the way? Ok, unplug it from both ends, then plug it back in." would be his first question. It was funny to us but I guess it's a real problem lol.

So earlier today I noticed that one of the gpu's in a 6 card 1070ti rig wasn't hashing.  I checked device manager in Windows 10 and it showed a Code 43 error.  I identified the problem gpu and started searching around online for soluctions.  Found a number of things ranging from a simiple restart to removing and reinstall the video drivers.  I did pretty much everything I could think of ...

Disabled and Re-enabled in device manager, uninstalled the gpu in device manager, I attempted to reboot in safe mode to use ddu, but my system would hang every time I tried to do that. (I later realized that I could get into safe mode just fine when I tried to do it with the gpu's unplugged)  At that point, I tried reverting back to a previous restore point, which failed several times.  So then I decided I would write my base Win10 image to the drive.  It's a pain, but not that big of a deal, since I made an image of every mining rig hdd after installing the OS and tweaking things exactly the way I wanted them, but prior to installing any gpu's.  At this point from what I was reading, I thought maybe it was a driver issue.  Tried different Nvidia drivers, nothing was working.

Still had the same Code 43 error in device manager on that one card.  I had swapped the riser earlier in the troubleshooting process with another in the same rig and I started thinking maybe I had a bad gpu.  So I have another 6 card gpu rig with identical hardware ... Board, processor, ram, gpu's are all the same.  I put the gpu into my other rig and it worked just fine. 

So back to my problem rig ... I put the known good working gpu from the other rig into the problem rig and I was right back to the Code 43 error.  Tried a brand new riser with new usb and power cable, same issue ... One thing I had not checked was the power from the card to the psu.  Earlier in the day I just assumed that it was getting power, if it was showing up in device manager.  So I check ... the cable pops right out of the psu when I moved it.  When I built the rig a few weeks ago, it must not have snapped in the cable on the psu end.  With the way everything is positioned in my rig frame, it is very hard to see what's what when you're looking at the cables plugged into the psu's.  Actually had to shine a flash light in there and still had a hard time seeing it. 

So after 6 hours of troubleshooting it ended up being one cable that had come unplugged from the psu.  I'm guessing the power from the riser was enough to get the gpu to show up in device manager and for some reason, that never occurred to me until I was 6 hours into it.

What a fun day.