Arch Linux

MikeDacre

Member

From: San Francisco Bay Area

Registered: 2013-01-18

Posts: 51

Website

USB Related Kernel Panic

On Friday I started getting the following error message in my log approximately 50 times per second:

usb-storage: Error in queuecommand_lck: us->srb = ffff88106f5d3c00

I left it for the weekend as I didn't have time to solve it, and it was still going 50 times a second on monday morning.

I unmounted all USB drives, which had no effect, then I went and disconnected all USB connected storage, which included 8 USB3 drives in RAID5 + LVM, 4 USB2 drives in RAID0, 1 USB2 ext4 drive, and 6 USB3 drives with various file systems connected via a plugable USB3 hub, which I think could be buggy. Yes, that is 19 external drives.

Finally, the log messages stopped, after disconnecting USB device number 2... unfortunately I don't know which USB device that was. I also immediately got the following message:

usb 1-3.1-port2: cannot disable (err = -71)

. Immediately after this I started getting kernel errors, particularly

general protection fault: 0000 [#1] PREEMPT SMP

,

WARNING: CPU: 6 PID: 500 at kernel/watchdog.c:265 watchdog_overflow_callback+0x9c/0xd0()

. Further down the error logs I started seeing nouveau in the call trace, not sure if that is related.

A few minutes later I started getting khugepaged soft lookup errors like this:

BUG: soft lockup - CPU#14 stuck for 22s! [khugepaged:135]

. This kept repeating for all 24 hyperthreaded cores (The system is dual 6 core CPUs hyperthreaded to 24 virtual cores) about once every two seconds. They system actually never panicked fully, but these soft lookup errors kept happening, I wasn't able to stop them, and the machine became almost completely unresponsive. I executed

reboot

and a reboot started, but 5 minutes later still had not completed, the consol just kept filling up with error traces and messages about CPUs being stuck for either 22 or 23 seconds. I eventually force powered the machine off and rebooted it. Seems to be running fine now....

One other potential confounder is that I has just installed crashplan pro e from aur and set it running.

Unfortunately, I am not great at reading kernel error messages. Is anyone here able to give me pointers on what may be going wrong?

The complete syslog for this period (minus the thousands of initial messages) is at: https://gist.github.com/MikeDacre/6a589e554e4ea476bf81

---

Home Page: www.michaeldacre.com
Lab: Hunter Fraser's Lab
GPG key: E76370D6

domentoi

Member

Registered: 2014-10-26

Posts: 5

Re: USB Related Kernel Panic

I have the same error:

kernel: usb-storage: Error in queuecommand_lck: us->srb = ffff8803c7b93680

For me, it first started on September 22. It only happens when I plug a certain external hard drive (doesn't matter which port, either. Tried it out with all the other ports) in. This external hard drive is 1.5 TB, from Toshiba. NTFS, if I remember correctly. I plug the ext hdd in, mount it...browse files...all is going well...until I try to unmount it from thunar. It lags, and leaving journalctl -f open shows that after a while the lines of the above show up. Just lines and lines of that error. I do not, however, have the other lines you do.

Plugging in smaller flash drives, mounting them, and unmounting works correctly.
Logfile of the end: https://gist.github.com/resurrexit/96f45482ba101d571f22