TrueNAS Scale Kernel Panic

TrueNAS General
1

Hi,

A couple of months ago my NAS went down, when I checked it, I noticed a kernel panic. After some troubleshooting (with some help from ChatGPT) I ran a memtest and found out one of my RAM sticks was bad. Replaced the faulty stick, re-ran memtest without errors and then reinstalled TrueNAS Scale and set everything up again. Was all fine until a couple days ago, when again I couldn’t reach my NAS. Important to note, at the time I was extracting 3x 50GB zip files on the NAS when it lost connection.
The message is
Kernal panic - not syncing: VERIFY3(rs_get_start(rs, rt) <= start) failed (23725884932096 <= 23725884915712)

This time I’m unfortunately not able to locate the problem. I assumed RAM again, because the extracting would be a RAM intensive operation. Ran memtest again, which did 4 passes without errors. Did some tests on the disks as well, which showed no errors. It seems to fail while trying to import the pool.

I will add some screenshots (apology for the blurry pictures), and hopefully someone can help me with what I should do next.

Hardware:
MSI B450M-A PRO MAX II
AMD Ryzen 5 4600G processor
Seasonic G12 GC-550 PSU
Cooler Master Silencio S400
G.Skill DDR4 Aegis 2x8GB 3200MHz
3x Seagate Ironwolf 8TB ST8000VN004 3,5 HDD SATA3 7200RPM
Gigabyte 256GB M.2 SSD (Bootdrive)

TruaNAS scale version:
25.04.1

Kernel panic:

20250907_230333
20250907_2303331920×1156 248 KB

Memtest results:

20250908_132731
20250908_1327312216×2243 371 KB

Some disk data (Smartctl), similar for all 3 disks:

20250908_154032
20250908_1540321920×1440 255 KB

Some more disk data + pool status

20250909_213435
20250909_2134351920×1479 204 KB

Tried to manually import pool:

20250909_220448
20250909_2204481920×1142 208 KB

3

Still could have been a RAM error. Just transitory.

That’s the frustrating thing with Non-ECC RAM, you can’t tell if it caused a problem, unless the problem is repeatable.

My tolerance of Non-ECC RAM is getting quite low. So if I can, any new computer will have ECC RAM. Though laptops are a *itch to try and get ECC RAM. The CPU in my HP ProBook supports ECC, (unlike some others), but nothing else, so I am out of luck…

1 Like
4

That’s a pity — your CPU (AMD Ryzen 5 4600G) actually supports ECC to some extent, but it looks like the motherboard doesn’t cooperate. If you had an ECC stick lying around, you could try it just to rule out RAM issues.
Otherwise, maybe try importing the pool on a different system to check if the problem is hardware-specific.

5

Searching on the old forum/reddit for the same error, has been suggested here and there to import the pool readonly zpool import -o readonly=true -f <pool>, and save much data is possible with a zfs send-receive.