Kernel panics and spinlock when RAID mirror mounted
#1
Hi everybody

I bought and setup my first NAS with two mirrored 4TB Seagate drives in April 2019.   Everything worked brilliantly and still does.
Code:
Linux pinky 4.4.167-1213-rockchip-ayufan-g34ae07687fce #1 SMP Tue Jun 18 20:44:4


Second NAS bought and configured with two mirrored 8TB Western Digital drives in January 2020; however kernel panics every 20 to 60 minutes when setting up a RAID mirror. 
  • No issue mounting and using the drives in a RAID stripe. 
  • Switched out the RockPro64 board, the SATA card (for identical Pine64 SATA card) and SD card; however same results. 
  • Tried multiple images and kernels and got the same results. 
My guess at that stage was the shucked drives or WD's handling of the RAID partition was causing the grief so bought two 8TB Seagate Barracudas.


With the Seagate drives, the panics where definitely less frequent.  Around 40% of the sync was complete before the first panic - after that I had to hard reset about half a dozen times to complete the eleven hour job.

With the RAID mirror unmounted, the device is stable and functions as normal.  When the RAID mirror is mounted, the kernel panic within four hours regardless of workload.

Code:
kernel:[  231.149029] Internal error: : 96000210 [#1] SMP
kernel:[  238.703488] BUG: spinlock lockup suspected on CPU#1, swapper/1/0
kernel:[  238.706806]  lock: 0xffffff800926fdf0, .magic: dead4ead, .owner: swapper/1/0, .owner_cpu: 1


Notes:
  • I am not sure if my first SATA card is an earlier generation than the two I subsequently bought.  Unfortunately I pulled the SATA connector away from the board on my original card when closing the case; it is currently held together with tape and is not in a state I want to switch it out for testing (one of the new ones was the replacement).
  • Everything was bought from Pine64 direct including the 12V/5A power supplies. 
  • Have tried multiple releases and am currently running 4.4.202-1237-rockchip-ayufan running but can test on any build if anyone is interested in further info.
Code:
Linux BRAIN 4.4.197-1236-rockchip-ayufan-g30faab37e339 #1 SMP Tue Oct 22 11:35:10 UTC 2019 aarch64 GNU/Linux


My next step is to get a PCIe SATA Marvell 88SE9230 and see if that resolves the issue for me; however if anyone else has any suggestions or tips to narrow the issue down further, they would be much appreciated.

Thanks!
#2
Hi,

some suggestions:
  • which type of cooling do you use in the system? The SMP lockup issues might be related to this, see https://forum.pine64.org/showthread.php?tid=5369
  • did you try to reproduce the issue with the psu from the working system? There have been some reports here on the forums about issues with the 5A psu shipped by Pine64 (another explanation could be that your new disks draw more power than the old ones, in this case it should fail with both psus).
  • which version of the board do you have? It seems like earlier versions have a resistor issue - see here: https://forum.pine64.org/showthread.php?tid=8374
#3
Good Evening,

If you get the chance, could you try a mainline kernel and see if it behaves better?


Possibly Related Threads…
Thread Author Replies Views Last Post
  RockPro64 has bad Memory (Software sogfaults and kernel panics) kop316 20 22,370 08-04-2021, 07:42 PM
Last Post: t4_4t
  zfs on rockpro64 - any dual nvme PCI adapter for software raid ? romadr 0 1,878 04-22-2021, 02:59 PM
Last Post: romadr
  Carte PCIe bloquant le démarrage du Noyau / PCIe card blocking Kernel boot dml-pc 4 4,839 02-17-2021, 11:35 AM
Last Post: dml-pc
  Sata Raid Marvell 4port 9235 stuartiannaylor 1 3,075 07-04-2019, 08:12 PM
Last Post: stuartiannaylor

Forum Jump:


Users browsing this thread: 1 Guest(s)