akimboflix

I’m experiencing frequent kernel panics and ATA errors on my Unraid server, and I need help diagnosing the root cause. Below are the details of my setup and the logs I’ve encountered: System Details: Motherboard: MSI PRO Z790-P WIFI DDR4 (MS-7E06) BIOS Version: 1.70 (07/13/2023) CPU: 13th Gen Intel® Core™ i9-13900K Unraid Version: 6.12.4 Error Logs: Kernel Panic and Invalid Opcode Errors: Logs frequently show invalid opcode: 0000 [#1] PREEMPT SMP NOPTI related to the md_mod module (unraidd process). Stack traces point to faults in md_mod during disk operations, causing the unraidd process to terminate unexpectedly. I've seen repeated kernel warnings and call traces indicating abnormal process exits. NVMe Drive Issues: I/O timeouts and aborted commands on the NVMe drive (nvme1), specifically errors like nvme nvme1: I/O Cmd QID timeout, aborting. These timeouts seem related to high disk utilization or potential hardware issues. ATA Errors: Logs show multiple ATA errors on ata3, including Emask 0x10 (ATA bus error) and irq_stat 0x08000000, interface fatal error. Errors include failed write commands (WRITE FPDMA QUEUED) and repeated hard resets of the SATA link. The drive on ata3 reconnects at 6.0 Gbps but the errors persist Are these errors indicative of a failing drive, SATA controller issues, or possibly deeper kernel-related bugs? Is there some newer gen intel BS going on? Having to do lots of unclean shutdowns, unfortunately. Some of my drives are reporting errors, I may just move out everything on Disk 2 and unplug it. I'm using hotswap bays so I'm gonna try plugging some of those drives in directly and see if I still have any problems. Thanks in advance for reading into my issues, please let me know if I need to provide any more info or run any tests rehoboam-diagnostics-20240829-1520.zip

I'm sure this has been requested before but is it possible to move the oldest/least-accessed files first?

Setting up a syslog server and flash but in the meantime, does anybody know what might have caused this? running 6.12.3, noticed it only started happening after changing my cpu

Maybe try doing this command and see if that creates the file? modprobe i915 Otherwise, I guess you could try creating it yourself. Add to the file 'options i915 force_probe=56a0' and reboot. In my system logs, I searched for 'force_probe' and it gave me my cpu code which I used to replace the 56a0 value. It didn't fully work at first but then I updated to 6.12.3 and everything started working

I updated to 6.12.3 and everything just worked! Hooray!

That seemed to reveal /dev/dri ! Thank you so much! However, it looks like there's still some issues. After adding the directory to plex, the server doesn't play any video on any device. Also, I noticed that the screen hooked up to the unraid server froze at fb0: switching to i915 from VGA. 😔Eventually this thing will work.. Can't shut down gracefully again

When I initially created a boot usb, the server would hang at i915 initialization on versions newer than 6.10.3. I wasn't able to access the terminal and had no experience with unRaid so I rolled back. Should I try updating? intel_gpu_top hasn't been working for me. GPU statistics doesn't show any info for my graphics either. Can you elaborate on how to force i915 in modprobe.d?

Been troubleshooting this problem and somehow got /dev/dri to show up after a reboot... but then the plex docker froze after adding the directory and I had to reboot again. Back to missing /dev/dri again lol. Setting up this server has been a game of tech wack-a-mole 🤪 EDIT: Having really weird issues now. Half the time I reboot, I see the screen freeze around the Samba settings but it still goes online. When it does, I can terminal into the /dev/dri folder on my other PC but the system gets all unstable and sometimes can't shut down properly. I've been getting freezes and all sorts of weird behavior. rehoboam-diagnostics-20230829-2024.zip

Trying to enable hardware transcoding by following this guide but the /dev/dri directory does not exist. I've tried modprobe i915 in the console with no luck. Pretty much ran out of forum posts to look at too EDIT: also looked through these instructions but I was not able to add /dev/dri to the PMS directory Plex dash says it's not hw transcoding and I believe iGPU should be enabled in the BIOS. I'm using an HDMI monitor connected to my mobo with no GPU. Oh, and yes, I have Plex Pass. Using an MSI Z790 motherboard. rehoboam-diagnostics-20230829-1758.zip

No, I can only see ports on the motherboard 1-6

I have an ASM 1166 but none of the hard drives I plug into it seem to show up. I have verified that they do work and I've tried all the slots on my MSI Z790 mobo. Is this just a faulty card or is there some configuring I need to do? rehoboam-diagnostics-20230828-2113.zip